Download Disordered eating and psychological help-seeking

Graduate Theses and Dissertations
Graduate College
2012
Disordered eating and psychological help-seeking:
The role of perceptions of need
LeAnn Mills Waterhouse
Iowa State University
Follow this and additional works at: http://lib.dr.iastate.edu/etd
Part of the Psychology Commons
Recommended Citation
Waterhouse, LeAnn Mills, "Disordered eating and psychological help-seeking: The role of perceptions of need" (2012). Graduate
Theses and Dissertations. Paper 12587.
This Dissertation is brought to you for free and open access by the Graduate College at Digital Repository @ Iowa State University. It has been accepted
for inclusion in Graduate Theses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For more
information, please contact [email protected].
1
Disordered eating and psychological help-seeking: The role of perceptions of need
by
LeAnn Mills Waterhouse
A dissertation submitted to the graduate faculty
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Major: Psychology (Counseling Psychology)
Program of Study Committee:
Lisa M. Larson, Major Professor
Daniel Russell
Norman Scott
David Vogel
Meifen Wei
Iowa State University
Ames, Iowa
2012
Copyright © LeAnn Mills Waterhouse, 2012. All rights reserved
ii
Table of Contents
List of Tables
iv
List of Figures
vi
Abstract
vii
Chapter 1. Introduction
1
Chapter 2. Review of the Literature
6
Chapter 3. Method
57
Chapter 4. Results
70
Chapter 5. Discussion
90
References
112
Tables
125
Appendix A
Eating Attitudes Test-26
139
Appendix B
Email Communications
140
Appendix C
SONA Posting Form
142
Appendix D
Informed Consent Form
143
Appendix E
Demographic Questions
149
Appendix F
Clinical Outcomes in Routine Evaluation – SF
147
Appendix G
Illness Perception Questionnaire – Revised
148
Appendix H
Sociocultural Attitudes Towards Appearance
Questionnaire
Decisional Balance Scale for Anorexia - Revised
153
157
Appendix K
Attitudes Towards Seeking Professional
Psychological Help
Help-Seeking Behavior Items
Appendix L
“Something Fishy” Homepage
159
Appendix I
Appendix J
155
158
iii
Appendix M
Student Counseling Service Homepage
160
Appendix N
Debriefing Form
161
Appendix O
Institutional Review Board Approval of Study
162
iv
List of Tables
Table 1
Descriptive Statistics for Complete Sample
125
Table 2
126
Table 5
Descriptive Statistics Split by Whether or Not Participant Meets
Clinical Criteria
Descriptive Statistics Split by Whether or Not Participant Wanted to
View Online Eating Issues Information
Descriptive Statistics Split by Whether or Not Participant Wanted to
be Directed to Schedule a Counseling Appointment
Correlations Among Demographics and Constructs
Table 6
Analyses with Constructs
130
Table 7
Hierarchical Multiple Regression Results with IPQ-R (insight) as a
Moderator
Binary Logistic Regression Model Results with IPQ-R (insight) as a
moderator
Binary Logistic Regression Variable Results predicting desire to
view online information about disordered eating with IPQ-R
(insight) as a Moderator
Binary Logistic Regression Variable Results predicting desire to
schedule a counseling appointment with IPQ-R (insight) as a
Moderator
Hierarchical Multiple Regression Results with SATAQ as a
Moderator
Binary Logistic Regression Model Results with SATAQ as a
Moderator
Binary Logistic Regression Variable Results predicting desire to
view online information about disordered eating with SATAQ as a
Moderator
Binary Logistic Regression Variable Results predicting desire to
schedule a counseling appointment with SATAQ as a Moderator
Hierarchical Multiple Regression Results with DBS-Benefits as a
Moderator
Binary Logistic Regression Model Results with DBS-Benefits as a
Moderator
Binary Logistic Regression Variable Results predicting desire to
view online information about disordered eating with DBS-Benefits
as a Moderator
131
Table 3
Table 4
Table 8
Table 9
Table 10
Table 11
Table 12
Table 13
Table 14
Table 15
Table 16
Table 17
127
128
129
131
131
132
132
132
133
133
133
134
134
v
Table 18
Table 19
Table 20
Table 21
Table 22
Table 23
Table 24
Table 25
Table 26
Table 27
Binary Logistic Regression Variable Results predicting desire to
schedule a counseling appointment with DBS-Benefits as a
Moderator
Regression Results with DBS-Burdens as a Moderator
134
Contributing Variables of Significant Regression Predicting
ATSPPHS
Binary Logistic Regression Model Results with DBS-Burdens as a
Moderator
Binary Logistic Regression Variable Results predicting desire to
view online information about disordered eating with DBS-Burdens
as a Moderator
Binary Logistic Regression Variable Results predicting desire to
schedule a counseling appointment with DBS-Burdens as a
Moderator
Regression Results with DBS- Functional Avoidance as a Moderator
135
Binary Logistic Regression Model Results with DBS-Functional
Avoidance as a Moderator
Binary Logistic Regression Variable Results predicting desire to
view online information about disordered eating with DBSFunctional Avoidance as a Moderator
Binary Logistic Regression Variable Results predicting desire to
schedule a counseling appointment with DBS-Functional Avoidance
as a Moderator
137
135
135
136
136
136
137
138
vi
List of Figures
Figure 1
Andersen and Newman’s model of help-seeking behavior (1973)
8
vii
Abstract
Although there has been a substantial amount of research done to examine general
psychological help seeking, very little of this research has focused on disordered eating.
Additionally, most research about disordered eating is conducted with women who are
already seeking help. The present study examines the impact of a woman’s perceived
need on her willingness to seek help. Participants are 249 women at a large Upper
Midwestern University who have reported mild to moderate eating issues. Multiple
hierarchical regression and binary logistic regression were used to test moderation
hypotheses on three measures of help-seeking (attitudes, desire to view online
information about disordered eating, desire to schedule a counseling appointment).
Moderation hypotheses were not supported, however results indicate that general
psychological distress, perceived drawbacks to the disorder, and perceived functional
avoidance did contribute to some models. Discussion emphasizes the need for a larger
sample size, more research with participants who have disordered eating but are not yet
in counseling, and importance of mean differences and correlations.
1
Chapter 1. Introduction
Eating disorders like Anorexia and Bulimia are experienced by a frighteningly
large percentage of college aged women. Research indicates that, at any one point, 5-15%
of college aged women have symptoms consistent with the diagnosis of an eating
disorder and an additional 20-35% report eating disorder traits that impact functioning
(Meyer, 2001; Meyer, 2005; Prouty, Protinsky, & Canady, 2002). Despite the substantial
number of college-aged women who struggle with disordered eating behaviors and
attitudes, few women seek help for these concerns (Cachelin, Rebeck, Veisel, & StriegelMoore, 2000; Meyer, 2001; Meyer, 2005). For example, Meyer (2005) sampled a group
of 294 college women and found that, of the 32 women reporting significant disordered
eating only 5 were currently in counseling for this concern. Additionally, of the
additional 74 women with mild to moderate disordered eating behavior and attitudes,
only 2 women were seeking therapy for this concern. In a recent 8-year longitudinal
study, Stice, Marti, Shaw, and Jaconis (2009) found that 12% of girls have some form of
eating disorder between the ages of 12 and 20
While the prevalence of the disorder alone is disturbing, the prevalence combined
with the impact of the disorder is even more so. In a meta-analysis of 42 published
articles, Sullivan (1995) reported a mortality rate of 5.6% per decade, or .56% per year.
This means that every year approximately one out of every 200 people diagnosed with
Anorexia Nervosa will die from the disorder. Even individuals who recover from an
eating disorder can have long-standing physical concerns including heart, liver, and
kidney problems, dental decay, irreversible bone loss, and mental health diagnoses.
2
There is a need for more research on eating disorders in general, and for eating
disorder help-seeking in particular. One of the first steps to helping someone with an
eating disorder is getting him or her in to treatment. Intervention and empirically
supported treatment studies are valuable, however the information these researchers
gather doesn’t become helpful until the individual struggling is actually seeking help.
Eating disorders are a relatively new diagnosis and were first put in the Diagnostic and
Statistical Manual of Mental Disorders III in 1980; the “newness” of the disorder likely
contributes to the shortage of research. When searching in PsychInfo for help seeking and
depression (“help-seeking,” “help seek” or “seek treatment” + “depression,” all in the
abstract) the search returns 415 peer reviewed journal articles. In contrast, when
searching for help seeking and eating disorders (“help-seeking,” “help seek” or “seek
treatment” + “eating disorder” or “disordered eating,” all in abstract) only 15 peer
reviewed journal articles are found. Clearly there is a large disparity.
To compound this problem, much of the existing research that can be applied to
help-seeking for an eating disorder includes only samples of women already in treatment
(Cachelin et al., 2000; Cachelin, Striegel-Moore, & Regan, 2006; Keel et al., 2002;
Goodwin & Fitzgibbon, 2002). Researchers sample groups of women who are currently
in treatment to understand what helped them seek treatment or what kept them out of
treatment in the past; only a few studies have sampled a group of women with eating
concerns who are not currently in treatment to find predictors of their help-seeking
attitudes, or intentions (Meyer, 2001; Meyer 2005).
There is also reason to question whether research on other mental health disorders
can generalize to eating disorders. Some mild disordered eating behavior is endorsed by
3
society (caloric or food group restriction, excessive exercise); this is not true in the same
way for most other mental health disorders. Additionally, some existing research
demonstrates that the experience of having an eating disorder is unique compared to other
mental health diagnoses. Specifically, research on the experience of benefits and
drawbacks of eating disorders has revealed a 3-factor solution (including benefits,
drawbacks, and the disorder’s function to avoid problems) where research with all other
disorders has revealed a 2-factor solution to this construct (benefits and drawbacks,
Cockell, Geller, & Linden, 2003). Clearly, the experience of having an eating disorder is
different from the experience of having other disorders; it seems possible, then, that the
choice to seek help or not seek help may also happen differently.
One model for help-seeking was developed in the 1970s by Anderson and
Newman (1973) to better understand medical help-seeking. These authors suggest that
three primary factors contribute to an individual’s choice to seek help: predisposing
characteristics, enabling resources, and need. When it comes to predicting help-seeking
behavior, we primarily have information about the predisposing characteristics (primarily
demographics; Parslow & Jorm, 2000; Wang et al., 2005a) and enabling resources
(finances, support, health insurance; Carragher et al., 2010; Harman et al., 2004; Young
et al., 2001; Vasiliadis et al., 2007); we don’t have as much information about how need
predicts help-seeking for a mental health concern.
According to Anderson and Newman’s theory, need (evaluated and perceived) is
an important predictive factor; if an individual does not have a problem (evaluated need)
or doesn’t believe they have a problem (perceived need) they will almost certainly not
attend counseling for a problem. For individuals with an eating disorder, perceived need
4
seems particularly salient. If an individual believes that their behavior is “normal” or
sanctioned by society, they will likely not perceive a need. Additionally, if an individual
is not feeling distressed by the drawbacks of their disorder (or is enjoying the benefits of
the disorder) they may not perceive a need for help. Finally, if an individual is aware that
they are in distress, but have not yet gained insight into the fact that aspects of their
distress (ex: depressed mood, avoidance of food, low self-esteem, obsessive thoughts) are
linked, it seems unlikely they will seek help.
Much of the existing literature on help-seeking that examined need as a predictor,
measured help-seeking behavior through a proxy, like intent, willingness, or attitudes
about help-seeking (Meyer, 2001; Meyer, 2005). There is a need for research that
examines behavior rather than self-report intent, willingness, or attitudes. The present
study aimed to address several of the needs in the literature. First, this study intended to
add to the literature on help-seeking behavior for an eating disorder; as opposed to selfreport willingness/intent/attitudes, individuals actually demonstrated a desire (or lack of
desire) to learn more about eating concerns or schedule a counseling appointment by
clicking on a link. Second, this study examined the impact of perceived need on the
relation between severity and help-seeking behavior. Individuals completed self-report
measures about several factors related to perceived need (internalization of societal
beauty norms, benefits and burdens of the disorder, insight into the fact that their varied
symptoms are related to a core eating problem). Third, this study sampled women who
were not in treatment and with varied eating concerns – ranging from fully fitting an
eating disorder diagnosis to mild/moderate problems with eating behaviors and body
image. In this way, the present study gathered more information about a frequently
5
targeted population - individuals who are not in treatment, but would likely benefit from
counseling. Information gathered from this study can provide the field with information
about who is and is not willing to seek help; it can also provide information that outreach
coordinators and counselors can use to adjust perceptions of need and encourage
counseling for eating concerns.
In the present study I hypothesized that the degree of eating concerns will predict
three forms of help seeking (attitudes toward psychological help seeking, desire to learn
about eating concerns and their treatment, and desire to learn about how to schedule a
therapy appointment). Additionally, I hypothesized that five aspects of perceived need
will moderate the relation between eating concerns and help seeking. Specifically, it was
hypothesized that a greater reported value of insight about their eating concerns, burdens
of the eating concerns, and awareness of the function the disorder serves to avoid
discomfort would be related to reporting more willingness to seek help while greater
reported benefits of their eating concerns and internalization of cultural beauty norms will
be related to reporting less willingness to seek help.
6
Chapter 2. Literature Review
The purpose of this literature review was to summarize the existing information
on help-seeking behavior as it applies to general mental health concerns as well as a
specific population of individuals at risk for an eating disorder and to summarize a
category of person-related factors (perception of the disorder) that may be particularly
impactful on help-seeking for individuals with disordered eating behavior/attitudes. There
are five sections addressing help seeking in general, eating disorder background and helpseeking, internalization of societal beauty norms, decisional balance, and cognitive
representations of the disorder.
Help Seeking in General
Many individuals who need professional psychological help never choose to seek
this help (Cachelin, Rebeck, Veisel, & Striegel-Moore, 2000; Meyer, 2001; Meyer,
2005). Researchers have begun to address this issue by exploring barriers and facilitative
factors to seeking help. Because gathering information on actual behavior is quite
challenging, much of the existing research predicts attitudes about help-seeking or
reported intent/willingness to seek help, as these two constructs are theorized to be
predictors of actual behavior. In this section of the review, I will explore predictors of
help-seeking behavior and attitudes/intent separately. Predictors of help-seeking can fall
into multiple categories. One model of help seeking suggests that help-seeking behavior
can be predicted with information from three categories: predisposing characteristics,
enabling resources, and need (Andersen, 1995, 1968).
Anderson’s behavioral model. To predict medical help-seeking, Andersen and
Newman (1973) proposed a basic model including three constructs: predisposing
7
characteristics, enabling resources, and need. This model has been widely used and, in
fact, applied to help-seeking for mental health at least as frequently as for physical health
(Andersen, 1995). Throughout the years the model has been expanded beyond the
original three predictors (grouped into “personal characteristics”) to include larger
external factors – like the overall healthcare system, and outcome factors – like
satisfaction and perceived health. Because the present study aims to predict initial helpseeking behavior (as opposed to more cyclical physical health help-seeking, as the most
recent model predicts), this review will focus on the original model.
The first predicting factor proposed by Andersen and Newman (1973) is
predisposing characteristics (see Figure 1). Three sub-categories comprised predisposing
characteristics including demographics (age, gender), social structure (education,
occupation, ethnicity), and health beliefs (attitudes, values, and knowledge regarding
health and healthcare). The author suggested that demographic characteristics predicted
medical help-seeking behavior because they could predict a person’s potential health
problems – older people are more likely to seek health care, as they are also more likely
to be ill. Predisposing characteristics discussed frequently in the mental health helpseeking literature are similar to those originally proposed in this medical model: age,
gender, race/ethnicity, and occupation as well as constructs not originally included such
as stigma of counseling, comfort with personal disclosure, and area of residence (rural vs.
urban).
The second factor, enabling resources, includes both personal and community
resources (Andersen and Newman, 1973, see Figure 1). In order to seek help, a helper
(hospital, counselor, treatment facility) must be available. Personal resources are also
8
theorized to predict help-seeking behavior. These originally included health insurance,
time off work, transportation, and money. Additional personal resources could include
Figure 1. Andersen and Newman’s model of help-seeking behavior (1973)
psychological strength (needed to admit a problem, find treatment options, etc), and
social support.
Need is the final predicting factor, and includes evaluated and perceived need
(Andersen and Newman, 1973, see Figure 1). Evaluated need seems an obvious predictor
– someone with no physical (or mental) health need ought not be seeking services.
Additionally, if they do originally seek services, but learn that their concerns are not
serious enough to warrant treatment, they will be unlikely to continue treatment.
Perceived need, on the other hand, is a more complicated factor. Whether objectively true
or not, if an individual does not believe he/she has problem, there would be no reason to
seek help. Many diverse factors can contribute to perceived need, especially in the case of
psychological help-seeking. Measured constructs have included severity of the disorder,
frequency of symptoms, level of functioning, benefits and burdens of the disorder, and
attribution of distress to an illness.
9
The remainder of this section will be divided into sections based on what is
predicted (behavior or attitudes/intent) and further divided by the three predictors
theorized by Andersen and Newman (1973).
Behavior. Much of the help-seeking behavior literature comes from large
epidemiological studies, where individuals are interviewed to determine any diagnoses
and then asked about help they have sought for these problems. In these studies, “helpseeking” only includes seeking professional help for a psychological concern. This can
include help from a general practice doctor, psychologist, psychiatrist, or other mental
health professional. “Help-seeking” does not include visiting a doctor for non-mentalhealth issues or talking with friends or family about one’s mental health concerns.
As an example, one such study, specifically looking at help-seeking for
depression, used latent class analysis to understand behavior (Carragher, Adamson,
Bunting, and McCann, 2010). With a sample of over 7,000 individuals with a lifetime
diagnosis of Major Depressive Disorder, the authors found that approximately one-third
of the sample received very little to no help for their diagnosis. Additionally, they found
that males were less likely to be in the partially- or highly-active treatment categories, as
were racial/ethnic minorities, individuals in lower socioeconomic groups, and individuals
without health insurance (Carragher, et al., 2010).
Predisposing characteristics. A good portion of the literature on help-seeking
behavior has focused on predisposing characteristics. Researchers have found that many
demographic variables including biological sex, age, marital status, and race/ethnicity
play a role in someone’s help-seeking behavior for a mental health concern (Carragher, et
al., 2010; Parslow & Jorm, 2000; Wang et al., 2005a; Wang et al., 2005b).
10
Biological Sex. Gender has been shown to predict whether or not an individual
seeks counseling. Women are more likely to seek counseling than men. In a survey of
10,641 adults, Andrews, Issakidis, and Carter (2001) found that women were 60% more
likely to use mental health services than men (odds ratio = 1.6). In eight out of an
additional nine epidemiological studies, sex was a significant predictor of whether or not
a person had sought help for a mental health disorder in the past twelve months
(Bergeron, Poirier, Fournier, Roberge, & Barrette, 2005; Bland, Newman, & Orn, 1997;
Carragher, et al., 2010; Parslow & Jorm, 2000; Rhodes, Goering, To, & Williams, 2002;
Sareen, Cox, Afifi, Yu, & Stein, 2005; Wang et al., 2005b; Young, Klap, Sherbourne, &
Wells, 2001). The disorders included for study typically were mood disorder, anxiety
disorder, and substance abuse. These nine studies included samples ranging in size from
1,092 to 125,000 adolescent and adult participants; most samples were approx 2,000 or
approx 10,000 participants. Authors of these studies reported odds ratios, logistical
regression, and and bivariate anlyses. One of the nine studies found no significant effect
of sex in a sample of 1,956 individuals with a mood disorder (Wang et al., 2005a).
Age. A second demographic variable predicting help-seeking that has been
explored is age. Researchers have found age to have a significant effect. However, there
is conflicting evidence about which age groups are most likely to seek help. Out of eight
studies, reported in seven articles reviewed here, seven studies reported a significant
effect of age (Bland et al., 2005; Carragher et al., 2010; Parslow & Jorm, 2002; Sareen et
al., 2005; Wang et al., 2005a; Wang et al., 2005b). One, however, reported that age only
a nearly significant effect (χ2 = 8.78, p = .07, Rhodes et al., 2002). Sample sizes ranged
11
from 1,964 to 125,493 adolescent and adult individuals. reported odds ratios, basic
logistical regression, Pearson’s correlations and bivariate anlyses.
There is inconsistency in the literature with regards to the relation of age and
help-seeking behavior. Some authors choose to examine age as a continuous variable
while others choose to create categories. Additionally, researchers who examine age as a
categorical variable rarely create the same categories. For example, one study breaks age
into two groups of people under age 45 and those 45 and older (Bland et al., 1997) while
another study divided the sample into quintiles (Rhodes et al., 2002). Of the seven studies
that found significant relations between age and help-seeking behavior, three authors
reported that “younger” people (either under age 45, or using age as a continuous
variable) were more likely to seek help than “older” people, defined as either over age 45,
over age 60, or using age as a continuous variable (Bland et al., 1997; Parslow & Jorm,
2000; Wang et al., 2005b) two authors found that individuals in the mid-range (age 3050) were more likely to seek help than those either younger or older (Sareen et al., 2005;
Wang et al., 2005a), and two authors reported that individuals age 46 or older were more
likely to seek help than individuals under age 46 (Carragher et al., 2010; Wang et al.,
2005a). While there is no distinct pattern to these inconsistent results, it does seem that in
samples of individuals with general mental health concerns, being younger was related
with a higher likelihood of seeking treatment (Parslow & Jorm, 2000; Wang et al.,
2005b), while in samples of individuals with a mood disorder, being older was related
with a higher likelihood of seeking treatment (Carragher et al., 2010; Wang et al., 2005a).
Two of these studies were reported in the same article, with different results
(Wang et al., 2005a). Interestingly, there was a significant effect of age for participants
12
with Bipolar disorder such that the oldest group (age 46+) was most likely to seek help,
however for participants with Major Depressive Disorder, individuals in the mid-age
group (age 26-45) were most likely to seek help (Wang et al., 2005a).
Marital Status. Research demonstrates that, in general, individuals who are
married are less likely to seek help than those who are not (Carragher, et al., 2010,
Jackson et al., 2007). In six samples, authors of five articles reported a significant relation
between marital status and mental-health help-seeking behavior (Bergeron et al., 2005;
Parslow & Jorm, 2000; Rhodes et al., 2002; Sareen et al., 2005; Wells, Robins, Bushnell,
Jarosz, & Oakley-Browne, 1994) and one reported no significant relation (Wang et al.,
2005a). Samples ranged from 1,092 to 125,493 adolescent and adult participants; most
samples were between 2,000 and 10,000 adult participants. Authors reported odds ratios,
basic logistical regression, and Pearson’s correlation statistics. Some researchers broke
their sample down beyond married and unmarried participants to include divorced,
separated, widowed, and co-habitating unmarried individuals.
Typically, authors reported that individuals who were divorced, separated, or
widowed were more likely to seek mental health help than individuals who are married
(Sareen et al., 2005). Two exceptions were found to the general empirical finding that
those who were living with a partner were less likely to seek help than those who were
living alone. The first, Wells and colleagues (1994), reported that individuals who were
divorced, separated, or co-habitating for more than one year were more likely to seek
help than individuals who were married or never married. In the second exception,
Parslow and Jorm (2000) found that divorced individuals were no more or less likely to
13
seek help than married individuals, although they found that individuals who were
separated or lived alone were more likely to have seen a psychologist.
Race and Ethnicity. In three epidemiologic studies, racial/ethnic minority
individuals (including African American, Latino, and Asian participants) are less likely to
seek professional help for a mental health problem than White individuals (Carragher et
al., 2010; Miranda & Cooper, 2004; Wang et al., 2005b). Wang and colleagues (2005b)
reported that, with a sample of 9,282 individuals, African American and Latino
individuals were approximately half as likely to use any form of mental health services
(odds ratio, OR = 0.5, 0.6 respectively). Examining a sample of 1,498 depressed adults
Miranda and Cooper (2004) reported that general practice doctors were equally likely to
recommend specialty mental health care to African American, Latino, and White
individuals, but that White and minority individuals did not seek specialized care equally.
Specifically, Latino and African American clients were less likely than White clients to
take anti-depressants when prescribed (odds ratio, OR = 0.3 and 0.6 respectively,
compared to White clients) and Latino clients were half as likely to seek specialty mental
health care (odds ratio, OR = 0.5). It was suggested that this might be due to cultural
values (rather than resources) such that Asian, Latino, and African American individuals
choose to speak with a family member, religious leader, or other community member
before “resorting” to seeing a professional helper.
Summary. Overall, research has typically demonstrated that predisposing
characteristics like sex, marital status, age, and racial/ethnic background have a
significant relation with help-seeking behavior. Specifically, women are more likely than
men to seek psychological help (Carragher, et al., 2010; Parslow & Jorm, 2000),
14
participants who were living with a partner were less likely than unmarried participants to
seek help (Carragher, et al., 2010, Jackson et al., 2007), and Caucasian participants were
more likely to seek help than African American, Latino, or Asian participants (Carragher
et al., 2010; Miranda & Cooper, 2004; Wang et al., 2005b). While age had a significant
relation with reported help-seeking behavior, the age group most likely to seek help
varied. A pattern that emerged suggests that in a sample of individuals with general
mental health concerns, being younger was related with a higher likelihood of seeking
treatment (Parslow & Jorm, 2000; Wang et al., 2005b), while in samples of individuals
with a mood disorder, being older was related with a higher likelihood of seeking
treatment (Carragher et al., 2010; Wang et al., 2005a).
Enabling resources. In addition to general predisposing characteristics –
primarily demographic variables – Andersen and Newman (1973) theorized that enabling
resources also impacted an individual’s choice to seek professional help. These include
financial resources (income, health insurance) as well as general resources (education,
time off work, etc). Focus has typically been on the relation of help-seeking behavior
with whether or not an individual has insurance, household income, and education level;
these constructs will be reviewed below. One challenging aspect of understanding the
relation between help-seeking and enabling resources is that there is substantial
intercorrelation between these constructs; individuals with more education frequently
have a higher household income and are more likely to hold a job that provides health
insurance.
Health insurance. Most existing research has found that whether or not an
individual has health insurance predicts whether or not they have sought help. Authors of
15
three out of four studies reported that individuals who had health care were more likely to
seek help than those without (Carragher et al., 2010; Harman, Edlund, & Fortney, 2004;
Young et al., 2001). Harman and colleagues (2004) examined type of insurance and
found that, in a sample of 1,347 depressed individuals, participants without insurance
were less likely to seek help than those with private insurance (odds ratio = .41) and that
individuals with Medicare were more likely to seek help than individuals with private
insurance (odds ratio = 2.02).
Using a sample of 451 depressed adults, one study reported no significant
difference in help-seeking behavior in individuals with and without health insurance
(Vasiliadis, Lesage, Adair, Wang, & Kessler, 2007) However, using their original sample
of 5,183 individuals, they reported that people without health insurance were twice as
likely to have a diagnosis of depression and were also significantly more likely to report
having an unmet need regarding psychological care (Vasiliadis et al., 2007). The authors
suggest that this relation is connected to the established relationship between
socioeconomic status and mental illness.
Educational attainment. Research generally demonstrates that individuals who
have more education are also more likely to seek professional psychological help. Five
out of six studies, published in five articles, reported that individuals with greater
education are significantly more likely to seek professional help for a psychological
problem (Carragher et al., 2010; Parslow & Jorm, 2000; ten Have et al., 2004; Young et
al., 2001; Wang et al., 2005a). Authors found no significant difference in one sample
(Wang et al, 2005a). Sample sizes ranged from 1,572 to 10,641 adult individuals and
typically considered diagnoses of substance abuse, anxiety, and mood disorders.
16
Researchers used a variety of statistical techniques (odds ratios, regression) and some
treated income as a categorical variable, whereas others treated it as a continuous
variable.
Interestingly, Wang and colleagues (2005a) found that, for individuals with Major
Depressive Disorder, educational attainment had no significant effect on help-seeking
behavior, but that for individuals with Bipolar disorder, individuals with more education
were more likely to seek help. In another study, the authors found that individuals with
greater education were more likely to see a psychologist, but that education had no
significant relation with use of a psychiatrist, general practice doctor, or other health
professional to address mental health concerns (Parslow & Jorm, 2000).
Income. The impact of income on help-seeking behavior is not clear; researchers
have found conflicting results. Two out of three studies found a significant relation
between income and help-seeking behavior (Carragher et al., 2010; Parslow & Jorm,
2000), however the relation in these studies is still tenuous. One study found no
significant effect (Wang et al., 2005b).
Carragher and colleagues (2010) found that individuals with a higher income
were more likely to be in the partially-active treatment group (likely had seen a therapist),
but that income had no significant relation with being in the highly-active treatment
group (i.e., very likely had seen a therapist, probably had been hospitalized and/or visited
an emergency room for mental health concerns). Specifically, they broke income up into
$20,000 increments and reported that, compared to individuals who make less than
$19,999 a year, all other income groups were more likely to be in the partially-active
treatment seeking category (odds ratios, OR = 1.33 – 1.96). In another study, authors
17
found that individuals with higher income were more likely to have seen a psychologist,
but that income had no relation with whether or not they had seen a psychiatrist, general
practice doctor, or any other health professional to address mental health concerns
(Parslow & Jorm, 2000).
Summary. Overall, research demonstrates that enabling resources have a
significant relation with psychological help seeking behavior. Specifically, a consistent
relation has been found between educational attainment and help seeking as well as
health insurance and help seeking, while the relation between income and help seeking
behavior is not as well established. Authors report that individuals with health insurance
are typically more likely than individuals without to have sought psychological treatment
(Carragher et al., 2010; Harman, Edlund, & Fortney, 2004; Young et al., 2001) and that
individuals with more education were more likely than individuals with less education to
have sought treatment (Carragher et al., 2010; Parslow & Jorm, 2000; ten Have et al.,
2004; Young et al., 2001; Wang et al., 2005a). Regarding income, research indicates that
there might be a relation with psychological help seeking behavior, but that this is
inconsistent across practitioners (psychologist, psychiatrist, general practice doctor) and
different degrees of help seeking (partially vs. highly active); one study found no
significant effect (Carragher et al., 2010; Parslow & Jorm, 2000; Wang et al., 2005b)
Need. Finally, Andersen and Newman (1973) theorized that evaluated need and
perceived need would predict an individual’s help-seeking behavior. Most large
epidemiological studies were unable to accurately assess an individual’s perceived need,
however these studies do sometimes assess for severity and comorbidy. These constructs
will be reviewed below. In much of the medical field (where Andersen originally
18
developed his theory), perceived need can often be assumed; an individual who just had a
stroke or who has the flu will almost certainly perceive a need for medical care. With
psychological disorders, however, this assumption becomes more risky. Little research
could be found examining the relation between perceived need and willingness to seek
professional psychological help. Research examining the connection between evaluated
need and help-seeking will be reviewed here.
Severity. Two studies reporting information about severity and help-seeking
behavior were found (Bland, Newman, & Ohn, 1997; Wang et al., 2005a). In both of
these, which were epidemiological studies of approximately 2,000 individuals, the
authors reported that people who experienced more severe mental illness were more
likely to have sought help. Specifically, Wang and colleagues (2005a) found that, in
individuals with a mood disorder, treatment-seeking was more likely if the participant
had a history of suicide attempts or reported that their mental illness severely interfered
with life roles.
Co-morbidity. Four studies were located where authors examined the relation
between co-morbidity and help-seeking for a mental health concern (Bergeron et al.,
2005; Bland, Newman, & Ohn, 1997; Young et al., 2001; Wang et al., 2005a). Authors of
all four studies reported a significant positive relation. Sample sizes ranged from 1,092 to
1,964 participants. Authors of two of the four studies reported more detailed information,
specifically that individuals with comorbid depression and anxiety were more likely to
seek professional mental health help than those with just one diagnosis (Young et al.,
2001) and that individuals who had a chronic medical condition were more likely to use
mental health services (Bergeron et al., 2005).
19
Summary. Overall, research indicates that degree of evaluated need is related to a
history of help-seeking behavior. Specifically, individuals reporting greater severity of
the disorder and/or comorbid disorders were more likely than those without severity or
comorbid disorders to seek help (Bergeron et al., 2005; Bland, Newman, & Ohn, 1997;
Young et al., 2001; Wang et al., 2005a).
Attitudes and Intent. While researchers are typically primarily interested in
help-seeking behavior, they often examine attitudes about therapy or intent to seek
therapy as a proxy. For this reason, much of what we know about more complicated
correlates of help-seeking is in the attitude/intent to seek help literature. Here I will
briefly summarize findings that are relevant to the present study including level of
distress, and anticipated risks/benefits of seeking help.
Level of distress. Four articles were found examining the relation between level
of distress and help-seeking attitudes or intent. Two articles reported a significant
positive relation (Hepworth & Paxton, 2007; Vogel & Wei, 2005) and two articles
reported a significant negative relation (Tishby et al., 2001; Wilson & Deane, 2010). In a
qualitative study, Hepworth & Paxton (2007) interviewed 63 women with past or present
bulimic behavior. The authors reported that degree of psychological distress was one of
the factors frequently mentioned by women when asked about what would (or did)
prompt them to seek help. Other factors mentioned included symptom severity,
interference with life roles, and health problems (Hepworth & Paxton, 2007).
Authors of a second study, sampling 355 college students, reported that distress
positively contributed to student help seeking intentions (Vogel & Wei, 2005).
Specifically, authors used structural equation modeling to examine the relationship
20
between attachment, distress, and help seeking and found that anxiously attached
individuals were more likely than avoidant attachment individuals to acknowledge
distress and seek help and that individuals with insecure attachment (both groups)
reported less social support than those with secure attachments, which contributed to their
distress and then to their help-seeking intentions (Vogel & Wei, 2005).
Conversely, authors of a third study reported that, in a sample of 1500 Israeli high
school students, there was a negative relation between level of distress and willingness to
seek help in one of the groups (Tishby et al., 2001). Specifically, authors reported that
there was no significant relation between level of distress and willingness to seek help in
the general group, but that in a subgroup of depressed and suicidal adolescents, there was
a negative relation between these two variables (Tishby et al., 2001). In the final article,
Wilson and Deane (2010) seek to explain this unexpected relation between level of
distress and willingness to seek help in depressed or suicidal individuals. The authors
sampled a group of 302 Australian University students and found that higher levels of
suicidal ideation were related to lower intentions to seek help from friends, family, or
professional providers. Additionally, they found moderation effects such that higher
levels of depression symptoms strengthen this negative relation (Wilson & Deane, 2010).
As depression involves a slowing-down of the body as well as feelings of hopelessness
and worthlessness, it may be that these depression symptoms have a unique effect on the
relation between level of distress and willingness to seek help.
Anticipated Outcomes. There are both positive and negative outcomes that may
be related to seeking professional help; the balance between the risks and benefits to
help-seeking has been found to contribute to an individual’s willingness to seek
21
counseling. Anticipated risk has been defined as an individual’s perception of danger in
sharing personal information with another person (Vogel & Wester, 2003). Two studies
were found examining the anticipated risks of therapy and help-seeking intentions (Deane
& Chamberlain, 1994; Deane & Todd, 1996). Authors of both studies reported that
individuals with higher fearlfulness of treatment were also less likely to express
willingness to seek treatment. Study methods were similar in both articles, however one
examined a sample of 263 college students (Deane & Chamberlain, 1994) and the other
examined a community sample of 107 with a mean age of 40 (Deane & Todd, 1996).
Authors measured distress using the Hopkins Symptom Checklist –21, measured
treatment expectations with the Psychotherapy Questionnaire, and measured help-seeking
with a single item about likeliness of seeking help from a psychologist if the individual
had personal problems (Deane & Chamberlain, 1994; Deane & Todd, 1996).
Researchers have also examined expected anticipated risks and benefits together
and found that both benefits and risks impact intentions and attitudes about therapy
(Vogel, Wester, Wei, & Boysen, 2005). Vogel and colleagues reported results from two
studies on this topic in one article. Authors administered measures of treatment fears,
anticipated risks and benefits of counseling, and intent to seek therapy to a group of 354
college students. Results indicated that both anticipated risks and anticipated benefits
significantly predicted intentions to seek help for interpersonal issues such that
individuals who perceived fewer risks and/or greater benefits were more likely to seek
help (Vogel et al., 2005). In the second study, authors surveyed 1,128 college students
with measures on anticipated risks and benefits as well as type of psychological concern.
22
Again, authors found that anticipated outcomes were associated with the decision to seek
professional psychological help (Vogel et al., 2005).
Eating Disorders
This section of the literature review will provide an overview of the relevant
eating disorder literature. Sections will include information about the prevalence and
course of eating problems, symptomology, diagnosis of the disorder, and help-seeking for
eating problems.
Prevalence and Course of Disorder. Eating disorders are quite prevalent. In fact,
it has been reported that 17.9% of women will, at some point in their lives, have a
diagnosable eating disorder (Kjelsas, Bjornstrom, & Gotestam, 2003). In a recent 8-year
longitudinal study, Stice, Marti, Shaw, and Jaconis (2009) found that 12% of girls will
have some form of eating disorder between the ages of 12 and 20. Additionally, several
studies including samples of undergraduate women have reported that 11%-17% of the
sample reports symptoms severe enough to suggest a diagnosis of an eating disorder
(Meyer, 2001; Meyer 2005; Prouty, 2002). Further, an additional 25%-33% of women
report some degree of disordered eating and/or body image concerns which may interfere
with their daily functioning (Meyer, 2001; Meyer 2005).
Although not all sub-threshold eating disorder behavior progresses to an eating
disorder, there is substantial evidence that symptoms increase and become more severe
over time (Shisslak, Crago, & Estes, 1995). Once a full diagnosis has been reached, many
people never fully recover. For Anorexia Nervosa (AN), it is estimated that 40% of
individuals diagnosed with the disorder recover fully, 30% show some improvement, and
20% experience a chronic course of illness (Steinhausen, Rauss-Mason, & Seidel, 1991).
23
Anorexia Nervosa also has a concerning mortality rate. In a meta-analysis of 42
published articles, Sullivan (1995) reported a mortality rate of 5.6% per decade, or .56%
per year. This means that every year approximately one out of every 200 people
diagnosed with AN will die from the disorder.
Symptomology of Eating Disorders. Currently, the Diagnostic and Statistical
Manual of Mental Disorders – 4th Edition , Text Revision provides diagnostic criteria for
three eating disorders: Anorexia Nervosa (AN), Bulimia Nervosa (BN), and Eating
Disorder Not Otherwise Specified (EDNOS, American Psychiatric Association [DSM-IVTR], 2000). The diagnosis of EDNOS is given to individuals with a variety of eating
concerns including Binge Eating Disorder (BED) as well as individuals with symptoms
consistent with AN or BN which do not yet meet full criteria for a diagnosis. Each of
these three diagnoses (AN, BN, and EDNOS) will be discussed individually below.
Anorexia Nervosa. The core symptom of Anorexia Nervosa (AN) is that the
individual refuses to maintain (or reach, if still growing) an acceptable weight specifically at least 85% of expected weight (American Psychiatric Association [DSMIV-TR], 2000). An individual must have a low body weight to meet criteria for AN. In
addition to a low body weight, a diagnosis of AN requires that the individual demonstrate
amenorrhea (no menstrual cycle for at least three months) as well as a strong fear of
gaining weight or getting fat. The final diagnostic criteria for AN is that the individual
report that their body size/shape has an unusually large impact on their self-evaluation,
that they deny the seriousness of their low body weight, or that they have a distorted
perception of their body size or shape. In addition to the general diagnosis of Anorexia
Nervosa, an individual receives one of two subtype diagnoses: restricting type or binge-
24
eating/purging type (American Psychiatric Association [DSM-IV-TR], 2000). This
subtype differentiates between individuals who are underweight and binge and purge and
those who only restrict their food intake.
Individuals with Anorexia Nervosa often present with additional psychological
and physical symptoms. Psychological issues can include depressive symptoms (e.g.,
depressed mood, withdrawing socially, feeling irritable, trouble sleeping, and decreased
interest in sex), obsessive-compulsive symptoms (e.g., preoccupation with thinking about
food, hoarding food or recipes), perfectionism, low self-esteem, constrained emotion and
need for control (American Psychiatric Association [DSM-IV-TR], 2000). These
individuals may also present with symptoms consistent with one or more personality
disorders. Individuals who fit the binge-eating/purging AN subtype are more likely than
those who fit the restricting subtype to report impulse control issues such as alcohol
abuse, to be sexually active, to have a history of suicide attempt(s), and to meet criteria
for Borderline Personality Disorder (American Psychiatric Association [DSM-IV-TR],
2000).
Physically, individuals with AN are often struggling; they are literally starving
themselves. Common physical concerns (examined with laboratory blood tests) include
anemia, dehydration, and problems with liver and heart functioning. Individuals may also
complain of abdominal pain, constipation, cold intolerance, low blood pressure, swelling
of the hands and feet, feeling lethargic, or feeling excess energy (American Psychiatric
Association [DSM-IV-TR], 2000).
Bulimia Nervosa. The core symptom of Bulimia Nervosa (BN) is binging and
purging (American Psychiatric Association [DSM-IV-TR], 2000). A diagnosis of BN
25
requires that an individual eat an unusually large amount of food in a constrained time
period, while feeling an inability to stop eating, and then perform some sort of
inappropriate compensatory behavior (vomiting, excessive exercise, use of
laxitives/diuretics, or fasting) at least twice a week for three months (American
Psychiatric Association [DSM-IV-TR], 2000). Additional symptoms include that the
individual’s views-of-self are largely impacted by their weight and/or body shape and
that the binge/compensatory behaviors do not occur exclusively during an episode of
Anorexia Nervosa. As in AN, there are two diagnostic subtypes for Bulimia Nervosa:
purging type and non-purging type. Purging involves a behavior to unnaturally remove
food from the body (e.g. self-induced vomiting or misusing laxatives, diuretics, or
enemas). Individuals who do not purge use other types of compensatory behaviors like
fasting or exercise. Individuals with BN are typically (but not exclusively) in the normal
weight range (American Psychiatric Association [DSM-IV-TR], 2000).
In addition to the core Eating Disorder symptoms, individuals with BN often
present with other mental and physical health problems. Psychological concerns include
mood disorders (Dysthymic Disorder or Major Depressive Disorder), Anxiety Disorders,
Substance Abuse/Dependence, and Borderline Personality Disorder (American
Psychiatric Association [DSM-IV-TR], 2000). While the related physical problems in
Anorexia Nervosa are nearly always related to the individual’s low body weight, physical
problems in Bulimia Nervosa are almost always related to purging. These potential health
problems include fluid and electrolyte imbalance, metabolic problems, loss of dental
enamel, calluses or scars on the hand from inducing vomiting, and cardiac and skeletal
26
problems if the individual uses syrup of ipecac to induce vomiting (American Psychiatric
Association [DSM-IV-TR], 2000).
Eating Disorder Not Otherwise Specified. As with most disorders in the DSMIV-TR, there is a Not Otherwise Specified category for individuals who have symptoms
of, but do not fully meet criteria for, an Eating Disorder (American Psychiatric
Association [DSM-IV-TR], 2000). The APA offers research criteria for one increasingly
common EDNOS, Binge-eating Disorder (BED). Diagnostic criteria are similar to those
for BN, however the individual must not use any compensatory behaviors and their
bingeing must occur twice a week for at least six months (as opposed to three months in
BN). Additional criteria state that the individual must report distress regarding their
eating and report specific characteristics of their bingeing behavior (ex: eating more
rapidly, until uncomfortably full, when not hungry, eating alone because of
embarrassment, and feeling depressed, guilty, or disgusted after a binge) (American
Psychiatric Association [DSM-IV-TR], 2000).
Additional examples of EDNOS are offered in the DSM-IV-TR, however it is
made clear that this is not an inclusive list. Several of these examples essentially require
that an individual meet all except one criteria for a full diagnosis (ex: meet all criteria for
AN except have continued menstruation; meet all criteria for BN except behavior
happens less frequently, an individual who uses inappropriate compensatory behavior but
does not binge). Other examples involve problem eating behaviors not included as
criteria for AN or BN, such as repeatedly chewing and spitting out – but not swallowing
– large amounts of food (American Psychiatric Association [DSM-IV-TR], 2000).
Individuals with a diagnosis of EDNOS are a very heterogeneous group.
27
Diagnosis and Measurement of Eating Disorders. As with most mental health
disorders, diagnosis of an eating disorder in a clinical setting frequently involves an
interview and a pen-and-paper or computerized assessment. The client may also be asked
to keep a record of his or her eating (for example, keeping a food journal to record
behavior and emotions), meet with a physician for a physical and lab work, and/or meet
with a nutritionist.
In a research setting, however, accurate diagnosis becomes more challenging,
primarily because the diagnostic interview is frequently time and cost prohibitive when a
large sample is called for. Assessment in research typically involves a self-report
questionnaire (with results reported on a continuum, categorically, or with a variety of
subscales), however some use a brief interview process. At least one example of each
type of self-report measure will be discussed below.
One of the most common assessments used, the Eating Attitudes Test (EAT-26,
Garner, Olmsted, Bohr, & Garfinkel, 1982) contains 26 items and results are shown on a
continuum. This measure is based on the original Eating Attitudes Test, which contained
40 items and was developed in 1979 by Garner & Garfinkel. In the EAT-26, participants
are asked to endorse items on a six-point Likert scale ranging from “never” to “always”
where responding to an item with the lowest three options (never, rarely, and sometimes)
earns a score of 0, and responding with the highest three option (often, usually, always)
earns 1-3 points, respectively. Research has demonstrated that a cut-off of 20 points most
reliably differentiates individuals with a diagnosable eating disorder from those not
meeting criteria for a diagnosis (Alvarez-Rayon et al., 2004; Garner, Olmstead, Bohr, &
Garfinkel, 1982; Mintz & O’Halloran, 2000).
28
The Questionnaire for Eating Disorder Diagnosis (Q-EDD, Mintz, O’Halloran,
Mulholland, & Schneider, 1997) is one example of a measure that provides categorical
results. In this self-report questionnaire, participants respond to 50 items. These are
scored using a flow chart that assigns each participant to one of three categories:
asymptomatic, symptomatic, or eating disordered. Individuals are assigned to the
asymptomatic group if they have reported no eating disorder symptoms, to the
symptomatic group if they have indicated some eating disorder symptoms but not enough
to meet DSM-IV eating disorder criteria, and to the eating disordered group if they report
significant symptoms consistent with a DSM-IV diagnosis. The authors reported
moderate test-retest reliability ranging from .54 to .64 over a one- to three-month period.
Interscorer agreement was 100% over 100 protocols (Mintz et al., 1997).
The Eating Disorder Diagnostic Scale (EDDS, Stice, Telch, & Rizvi, 2000) is a
second example of a measure that provides categorical results. Participants respond to 22
items, each of which asks about a specific diagnostic criterion for Anorexia Nervosa,
Bulimia Nervosa, or Binge Eating Disorder. For some items, individuals respond on a
six-point Likert scale (from “not at all” to “extremely”) and for other items, individuals
simply answer “yes” or “no.” The measure also includes open-ended questions about
height and weight. The authors provide scoring criteria, so that researchers can create
categories of participants (AN, BN, BED, and no diagnosis).
Unlike many self-report measures, the EDDS uses a scoring algorithm where a
diagnosis of Anorexia preempts a diagnosis of Bulimia, which preempts a diagnosis of
Binge Eating Disorder. Individuals who report a body mass index (BMI) less than 17.5,
based on weight and height report and report amenorrhea would be put in the Anorexia
29
group. Individuals who are on hormonal contraception that meet low weight criteria are
coded as amenorrheic, because hormonal contraception can result in a regular menstrual
cycle even for individuals who would not menstruating normally. Individuals who fit
criteria for the Bulimia group must report (a) regular eating binges marked by loss of
control and consumption of large amounts of food, (b) regular use of compensatory
behaviors, indictaed by a summed response of 8 or more on four items about
compensatory behaviors, and (c) undue influence of body weight or shape on selfevaluation. Individuals who reported (a) regular eating binges), (b) endorse at least three
specific criteria of binge eating such as eating more rapidly than normal or binging when
not hungry, (c) marked distress regarding eating and (d) absence of any compensatory
behaviors.
The Eating Disorder Examination (EDE, Fairburn & Cooper, 1993) is an
interview measure. Subjects are asked a variety of standardized questions designed to
make a differential diagnosis of an eating disorder based on DSM-IV criteria. Based on
factor analysis, there are four subscales including restraint, eating concern, shape
concern, and weight concern. The EDE has been shown to have good validity and
reliability. Aalpha coefficients for each subscale were acceptable (Restraint: .71-.75,
Eating Concern: .75-.90, Weight concern: .67-.70; and Shape concern: .70-.82; Byrne,
Allen, Lampard, Dove, & Furstland, 2009). Additionally, reported test-retest reliability
was satisfactory (>.70 over a 7-day period, .50-.88 over a 14-day period) for subscales.
Interrater correlations are typically greater than .90 (Byrne et al., 2009). Finally, eating
disordered individuals have been shown to score higher than non eating disordered
30
individuals on all subscales and eating disordered individuals who show improvements in
behavior also show improvements in subscale scores (Byrne et al., 2009).
Another measure that is commonly used both in research and clinical settings is
the Eating Disorder Inventory (EDI, Garner & Olmsted, 1984), which is currently in the
third edition. The EDI-3 (Garner, 2004) assesses for symptoms which are directly
connected to an eating disorder diagnosis (such as body dissatisfaction, drive for thinness,
and thoughts about binging/purging) as well as symptoms that are experienced in a range
of mental health concerns such as depression, anxiety, perfectionism, low self esteem,
and interpersonal insecurity. Participants respond on a four-point Likert scale and the
assessment includes 91 items and 12 subscales. When given as a complete measure, the
EDI-3 can help a clinician gain a big picture understanding of the client’s personality and
eating disorder symptomology. Researchers can choose to use just one or two of the EDspecific subscales to gather information about a participant.
Summary. There are a variety of measures with acceptable reliability and validity
that can be used to assess for disordered eating behavior, some of which provide
categorical information and some that provide results on a continuum. The Eating
Disorder Examination (EDE, Fairburn & Cooper, 1993) is a strong categorical measure
of disordered eating behavior, but can be cost- and time-prohibitive due to the need to
individually interview each participant. Stice, Telch, and Rizvi (2000) offer the Eating
Disorder Diagnostic Scale (EDDS), a self-report measure that provides similar
categorical results as the EDE, but requires much less time for the researcher. The Eating
Attitudes Test (EAT, Garner, Olmsted, Bohr, & Garfinkel, 1982), on the other hand,
assigns each participant a score on a continuum, with a cutoff above which most
31
individuals would fit an eating disorder diagnosis. The Eating Disorder Inventory (EDI
Garner & Olmsted, 1984) includes subscales that measure behavior and cognitions about
eating and exercise as well as subscales to measure personality traits often observed in
individuals with significant eating concerns. Certain subscales of this measure can be
used to assess severity of eating issues and others are very useful in treatment (EDI-3,
Garner, 2004). The present study will include the Eating Attitudes Test-26 because it
defines eating concerns on a continuum and does not assume severe pathology.
Help Seeking for Eating Disorders. While we have a plethora of data regarding
help-seeking behavior for other forms of mental health concerns (primarily related to
depression, anxiety, or substance abuse) we do not have large epidemiological studies to
help us understand correlates of help-seeking behavior for an eating disorder. We do,
however, have several studies that aim to understand a few constructs related to helpseeking behavior, most of which are correlational. In the longitudinal studies, all
participants have been identified as needing treatment for eating problems (and referred)
before the first contact. In this section, three factors related to help-seeking for an eating
disorder are included: ethnicity, comorbidity, and perceptions of need.
Ethnicity. Two studies were located that examined the relation of ethnicity and
treatment seeking behavior for an eating disorder (Cachelin, Rebeck, Veisel, & StriegelMoore, 2000; Cachelin, Striegel-Moore, & Regan, 2006). Authors of one study reported
no significant difference in ethnicity between the group who had sought treatment and the
group who had not (Cachelin, Rebeck, Veisel, & Striegel-Moore, 2000). The authors
administered the Eating Disorder Examination (EDE, Fairburn & Cooper, 1993) to a
group of community women and found 61 women who met criteria for an eating
32
disorder. As the sample was small and thus significance difficult to reach, it seems
important to report that the treatment-seeking group included twice as many Latina
women , two more African-American women, and one fewer Caucasian woman than the
non-treatment seeking group (Cachelin et al., 2000). The authors also reported no
significant differences between the treatment seeking and non-treatment seeking groups
on body mass index, socioeconomic status or insurance coverage, ethnic
identity/accultuaration, or specific ED diagnosis (Cachelin et al., 2000). The authors,
however, reported that individuals in the treatment group reported significantly more
distress related to bingeing, had earlier first experiences with overeating, and earlier onset
of regular overeating when compared to the non-treatment seeking group (Cachelin et al.,
2000). It was reported that, while 85% of the sample wanted treatment for an eating
problem, only 8% of the sample had received treatment for an eating disorder; an
additional 49% had received treatment for another concern (Cachelin et al., 2000).
In contrast, in a 2006 article it was reported that there was a significant effect of
ethnicity on treatment seeking for an eating disorder with a community sample of 190
women with eating disorders; 80 of these women were Mexican American and 110 were
European American (Cachelin, Striegel-Moore, & Regan). The authors used the
Structured Clinical Interview for DSM-IV-TR (SCID-IV-TR; First, Spitzer, Gibbon, &
Williams, 2001) to assess for any Axis I disorder; participants who met criteria for an
Axis I disorder were then administered the Eating Disorder Examination, 12th edition,
which provided a specific eating disorder diagnosis. Using logistic regression, the authors
found that a model including SES, ethnicity, age of onset, presence of psychiatric
comorbidity before onset, and type of disorder (AN, BN, BED) significantly
33
distinguished between individuals who had and had not sought treatment. Ethnicity was
the strongest predictor (z=8.14, p <.001) such that European American women were 2.74
times more likely to have sought treatment than Mexican American women (Cachelin,
Striegel-Moore, & Regan, 2006).
The authors also examined the impact of acculturation on treatment-seeking
behavior. Because only Mexican American women were given the acculturation
assessment, a second logistic regression was performed where the ethnicity variable was
replaced by acculturation. This model of seven predictors was nearly significant, χ 2(7) =
13.54, p =.06. While the model did not reach significance, authors reported that
acculturation contributed significantly (z=4.1, p = .04) such that women with a higher
degree of acculturation were 1.75 times more likely to have sought treatment than women
with a lower degree of acculturation (Cachelin, Striegel-Moore, Regan, 2006). The
authors also reported that individuals with Bulimia and more frequent purging, longer
duration of the disorder, and psychiatric comorbidity were more likely to be in the
treatment-seeking group than individuals with any other diagnosis and fewer symptoms
(Cachelin, Striegel-Moore, Regan, 2006).
Comorbidity. Two studies were found that directly examined the relation of
comorbidity and treatment-seeking (Goodwin & Fitzgibbon, 2002; Keel et al., 2002). In
one article, it was reported that individuals with more severe pathology and comorbidity
were more likely to seek treatment. Using a longitudinal method (interview to diagnose
and then every six months for five years) and a sample of 246 individuals diagnosed with
AN or BN, the authors reported that over 95% of the sample had received some form of
ED treatment (Keel et al., 2002). Additionally, the authors reported that individuals with
34
comorbid mood disorders, personality disorders, and/or a lower Global Assessment of
Functioning (GAF) score were more likely to have received treatment for a longer period
of time; individuals with Anorexia Nervosa were more likely than those with Bulimia
Nervosa to have received inpatient treatment (Keel et al., 2002).
Authors of a second study specifically examined the impact of social anxiety on
help-seeking for an eating disorder (Goodwin & Fitzgibbon, 2002). The authors
administered assessments (the EDI, Garner & Olmsted, 1984; Bell Object Relations
Inventory, BORI; Bell, Billington, & Becker (1986) to a group of 28 individuals
presenting at an outpatient eating disorder clinic for a first assessment appointment.
Researchers compared two groups of individuals: those who had returned for their
recommended first therapy appointment and those who never returned to the clinic. They
found that individuals who did not engage in treatment had significantly higher scores on
the Insecure Attachment subscale (measuring social anxiety about being liked, fears and
sensitivity to rejection, easily hurt by others) of the Bell Object Relations Inventory (F=
8.29, df = 1, p < .05), but that there were no other significant differences between these
groups including diagnosis or severity. While it seems that comorbidity can increase an
individual’s willingness to seek treatment, in some circumstances certain
psychopathology can also decrease an individual’s ability to seek help.
Perception of need. Two studies were found which examined an individual’s
perceived need for therapy (Meyer, 2001; Meyer, 2005). In the first study, Meyer
reported that individuals with substantial disordered eating behavior who were not
currently in treatment attributed their choice to not seek treatment to a belief that their
behaviors did not warrant therapy (2001). This study, which sampled 248 college
35
students, aimed to examine the relation between attitudes toward seeking help, eating
disorder diagnoses, and reasons for not seeking treatment. Meyer (2001) determined that
38 individuals were eating disordered (defined as meeting diagnostic criteria for an eating
disorder), an additional 78 were symptomatic (defined as reporting some eating disorder
symptoms), and 122 were asymptomatic (defined as not reporting notable eating disorder
symptoms).
Help-seeking behavior was assessed with three yes/no items regarding (1)
whether they believed they needed to seek professional counseling for their eating
concerns, (2) whether they had received counseling for these concerns in the past, and (3)
whether they were currently seeing a therapist for these concerns. Two participants (in
the symptomatic group) reported that they were currently in counseling and two
additional participants (also in the symptomatic group) reported a history of treatment.
No one in the eating disordered group reported current or past treatment. When asked
about whether they believed they needed treatment, 23% of the eating disorder group and
6% of the symptomatic group said “yes,” 37% of the eating disorder group and 28% of
the symptomatic group said “maybe,” and 40% of the eating disorder group and 66% of
the symptomatic group said “no” (Meyer, 2001).
To assess general attitudes toward seeking professional psychological help,
participants were administered the Attitudes Toward Seeking Professional Psychological
Help – Short form (ATTSPPH-SF), a 10-item self-report measure of attitudes toward
traditional counseling. Individuals in the diagnostic categories (eating disordered,
symptomatic, asymptomatic) reported statistically equivalent attitudes on this measure
36
(Meyer, 2001). The author did not report analyses including the participants’ perceived
need for counseling and their attitudes toward counseling.
Finally, Meyer (2001) assessed reasons why participants were not currently in
counseling. For participants in the eating disordered category, most common responses
were “don’t want anyone to know” (40%), “problem not worrisome enough to me”
(32%), and “don’t believe I have a problem at all” (24%). Individuals in the symptomatic
group reported similar reasons including “problem not worrisome enough to me”(50%),
“don’t believe I have a problem at all” (35%) and “don’t want anyone to know (21%).
Meyer also reported that participants were least likely to cite “fear of therapy” or “lack of
awareness of help available” as reasons for not attending counseling (Meyer, 2001).
In her 2005 study, Meyer aimed to understand the impact of treatment anxiety,
psychological defenses, and agreement with societal norms of beauty on help seeking for
eating concerns. The Q-EDD was administered to 294 undergraduate college women;
results of the assessment indicated that 11% of the sample (32 participants) met criteria
for the eating disordered group, an additional 25% (74 participants) met criteria for the
symptomatic group, and 64% (188 participants) were categorized in the asymptomatic
group. Help-seeking behavior was assessed with two yes/no questions (1) “are you
currently seeing a counselor/psychotherapist for your dieting/body image concerns or
eating disturbance?” and (2) “Do you feel that you need counseling or therapy for your
dieting habits, body image concerns, or eating disturbance?” (Meyer, 2005).
Five of the 32 eating disordered participants and two of the 74 participants in the
symptomatic group reported currently being in therapy for eating concerns. The author
determined, through Fisher’s Exact Test, that individuals in the eating disordered group
37
were significantly (p < .05) more likely than those in the symptomatic group to be in
therapy (Meyer, 2005). Additionally, the participants’ belief that they need therapy for
eating concerns was examined. It was found that 56% of the eating disordered group and
39% of the symptomatic group believed they needed treatment for their eating concerns.
Conversely, 44% of the eating disordered group and 61% of the symptomatic group
denied that they needed treatment for their eating problems (Meyer, 2005). The
percentage of participants in each group that believed they need treatment was noticeably
different in this study compared to the prior study (Meyer, 2001). Some of this disparity
may be due to the fact that in one study participants were given a “maybe” option about
their need for treatment and in the other study were only given “yes” or “no” options.
In addition to degree/type of eating disorder and treatment seeking beliefs, Meyer
(2005) assessed the participants’ degree of immature defenses (using the Defense Style
Questionnaire, DSQ-40; Andrews, Singh, & Bond, 1993), their thoughts about therapy
(using the 19-item Thoughts About Psychotherapy Scale, TAPS; Kushner & Sher, 1989),
and the degree to which they endorse societal beauty norms (using the Beliefs About
Attractiveness Scale-Revised, BAA-R; Petrie, Rogers, Johnson, & Diehl, 1996). Meyer
analyzed differences between the groups’ scores on these three scales (2005). She
reported significant findings of a MANOVA (Wilks’s Lambda = .90, F(3, 291) = 5.45, p
= . 0001). Two follow-up ANOVAs were significant. First, it was reported that the
symptomatic group scored significantly higher than the asymptomatic group on the
Immature Style Defense Scale of the DSQ-40; the ED group mean on this scale was
between the asymptomatic and symptomatic group’s means and was not significantly
different from either group. Second, it was reported that individuals in the symptomatic
38
and eating disordered groups endorsed beliefs that individuals must be thin, fit, and
attractive to be successful, valuable, and/or happy significantly more strongly than
individuals in the asymptomatic category (Meyer, 2005).
The author also reported that participants who indicated that they did not believe
they needed treatment were also less likely to endorse societal beauty norms; this finding
was the opposite of what was hypothesized (Meyer, 2005). To better understand this
finding, further analyses were run. It was found that individuals who indicated they did
not believe they needed treatment reported significantly higher levels of immature
defenses. Meyer suggested that perhaps these individuals use immature defenses (such as
denial) to down-play the degree to which their own body image is impacted by their
endorsement of societal beauty norms (Meyer, 2005).
Summary. Overall, both studies indicate that an individual’s perception of their
need for treatment is often inconsistent with an objective, clinical assessment of their
need for treatment (Meyer 2001; Meyer 2005). Many participants in the eating disordered
and symptomatic groups indicated that they did not believe their behaviors and attitudes
about food and their body warranted therapy. Additionally, it was reported that
individuals who indicated they did not need treatment were less likely to strongly endorse
societal beauty norms and were more likely to report higher levels of immature defenses
(Meyer, 2005).
Internalization of Thin-Ideal
Western society sends relatively strong, explicit messages about what is and is not
beautiful; for women, small body size is often seen as a requirement for beauty. Some
researchers have hypothesized that internalization of these societal messages may
39
contribute to development of an eating disorder. In this section of the literature review, I
will discuss the construct of internalization and related theory as well as discuss relevant
research involving the construct.
Internalization and related theory. A common belief about eating disorders is
that individuals begin negative eating behavior, at least in part, to become thinner, and
that these individuals want to be thin in order to measure up to a societal standard of
beauty. In this way, a high degree of internalization of the thin-ideal can strengthen other
negative constructs related to eating disorders including negative affect, body image
dissatisfaction, and dieting (Stice, 2001).
Measurement of internalization of the thin-ideal began near the end of the last
century, when the Sociocultural Attitudes Towards Appearance Questionnaire (SATAQ)
was developed (Heinberg, Thompson, & Stormer, 1995; Stice, Schupak-Neuberg, Shaw,
& Stein, 1994). In addition to measuring internalization of societal beauty norms, some
questionnaires also measure awareness of societal beauty norms. These two constructs
are different in that awareness suggests that the individual can label current norms of
beauty and may or may not agree with these norms while internalization suggests that the
individual has taken these beliefs on as his or her own self-standards.
Several theories have incorporated thin-ideal internalization as one causal factor
in the development of an eating disorder. One of these, social reinforcement theory, states
that individuals internalize beliefs that are held by significant others (Kandel, 1980).
According to this theory, family, friends, and the media have provided reinforcement for
the thin-ideal through words or actions (teasing, criticism, avoidance, praising dieting,
40
etc) that clearly communicate the benefits of thinness, such as acceptance, happiness, and
beauty (Hohlstein, Smith, & Atlas, 1998).
Stice and Agras (1998) offer the sociocultural or dual-pathway model. In this
theory, the authors suggest that thin-ideal internalization leads to negative emotionality
and dieting, which increase the risk for eating disorder behavior – particularly bulimic
behavior. Research support for this theory will be discussed in the next section.
A third theory including internalization of societal beauty norms is objectification
theory, originally developed by Fredrickson and Roberts (1997). In this theory, the
combination of thin-ideal internalization and an awareness of sexual objectification lead
an individual to habitually monitor their body (frequently looking in mirrors, scrutinizing
particular body parts or weight, etc.), which leads to body dissatisfaction, then negative
emotionality (anxiety and depression) and finally to disordered eating behavior (Mitchell
& Mazzeo, 2009). In all three of these theories, internalization of the thin-ideal is posited
to be a factor in how an individual perceives him/herself and his/her value to society.
Empirical Support. Recently, the construct of thin-ideal internalization has
gained more attention in the literature. In fact, Thompson and Stice - both very prolific
researchers in the realm of eating disorders – have clearly stated the importance of thinideal internalization and encouraged others in the field to increase research of this
construct (2001). While we do have a substantial amount of information regarding the
correlates of thin-ideal internalization, only one article was found that examined the
connection between thin-ideal internalization and help-seeking for an eating disorder
(Meyer, 2005, discussed above). Research on thin-ideal internalization has primarily
41
focused on body image dissatisfaction and eating disordered behavior; findings of this
research are reviewed below.
Constructs predicted by thin-ideal internalization
Body image dissatisfaction. Eight articles, containing ten samples, were found
that reported correlational analyses of thin-ideal internalization and body image
dissatisfaction (Bedford & Johnson, 2006; Forbes, Jobe, & Revak, 2006; Griffiths et al.,
1999; Low et al., 2003; Morrison & Sheahan, 2009; Rukavina & Pokrajac-Bulian, 2006;
Tiggemann, 2003; and Vartanian & Hopkinson, 2010).
Sample sizes ranged from 49 to 300 participants and included only females,
ranging from high school age to a sample of elderly women (age 70 +/- 3yrs). Authors of
all eight studies measured thin-ideal internalization using the SATAQ (Heinberg,
Thompson, & Stormer, 1995). Authors of half of the articles reported measuring body
dissatisfaction with two subscales of the Eating Disorder Inventory (EDI, Garner &
Olmsted, 1984), the Drive for Thinness and the Body Dissatisfaction subscales (Griffiths
et al., 1999; Low et al., 2003; Morrison & Sheahan, 2009; Vartanian & Hopkinson,
2010). Measurement of body dissatisfaction in the other four articles was not consistent,
but included using the difference between perceived current and ideal body types, the
Body Esteem Scale (BES, Mendelson, Mendelson, & Andrews, 2001), and the Body
Shame subscale of the Objectified Body Consciousness Scale (OBCS; McKinley &
Hyde, 1996).
In nine out of the ten samples, authors reported positive significant correlations
between thin-ideal internalization and body image dissatisfaction. The one exception was
reported by authors who tested two samples of women and measured thin-ideal
42
internalization and body image satisfaction (rather than dissatisfaction); they found that
in a sample of “older” women (age 67-73), there was no significant correlation (r = -.24,
p> .05), but that there was a significant correlation (r = -.52, p < .01) for college-aged
women (Bedford & Johnson, 2006). The lack of significant findings may be due to
insufficient power. In studies using the EDI, authors reported consistent, positive
correlations between the SATAQ and the Drive for Thinness subscale (rs = .42-.58,p <
.01) and the SATAQ and the Body Dissatisfaction subscale (rs = .37-.50, p < .01). In one
study, authors combined these two subscales and found an even larger correlation (r =
.68, p < .001).
Reported results in other studies also suggested a significant relation between
body image dissatisfaction and thin-ideal internalization. In one study, authors reported a
significant correlation between the body shame subscale of the OBCS and the SATAQ
pressure subscale (r = .34, p <.001) and the SATAQ general subscale (r = .66, p < .001).
These same authors reported significant correlations between the SATAQ general
subscale and desired changes in body size: larger bust size (r = .34, p <.05), greater
height (r =.23 , p < .05), decreased weight (r = -.33, p < .05), and decreased lower body
size (r = -.27, p < .05, Forbes, Jobe, & Revak, 2006). In another study, authors reported a
significant correlation between SATAQ scores and appearance satisfaction (r = -.47, p <
.001) and weight satisfaction (r = -.32, p < .001, Rukavina & Pokrajac-Bulian, 2006).
Eating Disorder Behavior. Four articles, containing five samples, were found that
examined the relation between thin-ideal internalization and eating disorder behavior
(Griffiths et al., 1999; Griffiths, et al., 2000; Rukavina & Pokrajac-Bulian, 2006; and
Vartanian & Hopkinson, 2010). Authors of all four studies measured thin-ideal
43
internalization with the SATAQ (Heinberg, Thompson, & Stormer, 1995). The Eating
Attitudes Test (EAT, Garner et al., 1982; Garner & Garfinkel, 1979) was used to measure
eating behavior in two articles (Griffiths et al., 1999 and Rukavina & Pokrajac-Bulian,
2006) while the Revised Restraint Scale (RRS, Herman, Polivy, Pliner, Threkheld, &
Munic, 1978), a 10-item measure of dietary restraint, was used in the other two studies
(Griffiths et al., 2000 and Vartanian & Hopkinson, 2010). Sample size in these four
studies ranged from 82 to 300 participants; all four studies included only females.
Additionally, two studies sampled only college students, one sampled only high school
girls, and another sampled women aged 17-30.
Results in these four articles were consistent. Authors of all four articles reported
that there was a significant, and often large, correlation between thin-ideal internalization
and eating disorder behavior, such that as thin-ideal internalization increased, so did
eating disorder behavior. There was one exception: Griffiths and colleagues (1999)
examined this relation in both eating disordered and non-eating disordered individuals.
There was a significant correlation between thin-ideal internalization and eating disorder
behavior for individuals with an eating disorder (r = .41, p < .001), but this correlation
was not significant for individuals without an eating disorder. In the other three articles,
testing a general sample of participants (with and without ED symptoms), reported
correlations were positive and significant (range r = .45 -.61, p <.001).
Predictor of thin-ideal internalization: Media exposure. Two articles were found
that examined the relation between media exposure and thin-ideal internalization (Morry
& Staska, 2001; Tiggemann, 2003). Several additional articles examining the relation
44
between media exposure and body image dissatisfaction were found, however they will
not be reviewed here because they do not explicitly include thin-ideal internalization.
In the first study, authors aimed to examine the relations between magazine
exposure, thin-ideal internalization, self-objectification, eating attitudes, and body
satisfaction. They sampled 150 college students (61 male, 89 female) in an Introductory
Psychology course (Morry & Staska, 2001). Authors measured eating disorder
symptomology with the Eating Attitudes Test (EAT-40, Garner & Garfinkel, 1979), a 40item self-report questionnaire measuring behaviors, attitudes, and emotions about food
and body. Additionally, the authors measured thin-ideal internalization with the
Sociocultural Attitudes Towards Appearance Questionnaire (SATAQ, Heinberg,
Thompson, & Stormer, 1995). Authors also measured body concerns/feelings of fatness,
exposure to magazines in the past month, and self-objectification (Morry & Staska,
2001). Results of regression analyses indicate that magazine exposure predicted degree of
thin-ideal internalization. Additionally, authors reported that thin-ideal internalization
mediated the relationship between magazine exposure and body-shape dissatisfaction
(Morry & Staska, 2001).
In the second study, the author investigated the impact of television and
magazines on thin-ideal internalization, body dissatisfaction, and disordered eating and
found that magazines and television have different effects on thin-ideal internalization
(Tiggemann, 2003). Tiggeman administered several measures to a sample of 104
Australian female college students including media exposure, body dissatisfaction, selfesteem, and Body Mass Index. Additionally, the Eating Disorder Inventory (EDI, Garner,
45
Olmstead, & Polivy, 1983) and the SATAQ (Heinberg, Thompson, & Stormer, 19995)
were administered.
Correlational results indicated that both magazine and television exposure were
related to body dissatisfaction (r = .19, .17 respectively, p < .05), but that only magazine
reading was significantly correlated with thin-ideal internalization (r = .31, p < .01).
Amount of time spent watching TV was not significantly correlated with SATAQ scores,
indicating that it’s impact on body dissatisfaction occurs through a different mechanism
(Tiggemann, 2003).
Decisional Balance
In the transtheoretical model (TTM), Prochaska and DiClimente (1985) suggest
that people move through different stages when making behavioral change, and that these
stage movements are influenced by the individual’s decisional balance. Decisional
balance refers to the balance of benefits and burdens of an individual’s current problem.
As an example, this model was used to understand smoking cessation – an individual
who has no desire to change would likely believe that the benefits to smoking outweigh
the drawbacks; once this balance shifts to the drawbacks outweighing the benefits,
individuals may begin to consider – or initiate – change. This same theory has been
applied to a multitude of health behaviors such as weight loss, contraceptive use, and
cancer screening (Grimley, Riley, Bellis, & Prochaska, 1993; O’Connell & Velicer, 1988;
Velicer, DiClimente, Prochaska, & Brandenburg, 1985; Rimer, Conaway, Lyna, &
Rakowski, 1996) Additionally, an attempt has been made to understand decisional
balance in eating disorders (Cockell, Geller, & Linden, 2002).
46
Decisional Balance and Eating Disorders. One article was located that examined
decisional balance in eating disorders as it related to Prochaska and DiClimente’s stages
of change (Cockell, Geller, & Linden, 2002). The authors sampled 246 anorexic women
at various stages of treatment from three different eating disorder treatment programs.
Before participating, all of these women either self-identified as having Anorexia or were
in treatment and were identified by a clinician. Sixty-six percent of the women currently
met criteria for Anorexia, 29% met criteria for sub-threshold Anorexia, and 3% had met
criteria for Anorexia in the past (Cockell, Geller, & Linden, 2002). Participants
completed demographic information and a decisional balance measure developed by the
authors. This measure included 60 items (30 pro and 30 con items).
The primary purpose of the study was to develop this measure; the details of this
process are discussed in the methods section of the current paper. Most interestingly,
using exploratory factor analysis, the authors determined that there were three factors in
decisional balance for Anorexia – benefits, burdens, and functional avoidance. As
previous research of health behaviors – like smoking cessation and weight loss – had
determined there were only two factors, the appearance of three factors was unexpected.
After examination of the items, authors labeled the third factor ‘functional avoidance,’
because all items dealt with Anorexia serving the purpose of avoiding difficult issues,
like planning for the future, sexuality, relationships, and painful emotions (Cockell,
Geller, & Linden, 2002).
Decisional Balance and Help-Seeking. No research could be found that
examined decisional balance and help seeking; instead, most research with any decisional
balance scale purports to predict stages of change. It seems likely that moving from an
47
early stage of change (precontemplation, where an individual is unaware of a need for
change) to a later stage of change (contemplating change or even acting to change) would
also be correlated with an individual’s decision to seek help, however we don’t have
empirical evidence of this. What we do know, however, is that decisional balance does
predict stages of change. Specifically, as an individual moves from the precontemplation
to contemplation stage their awareness of burdens of the disorder rises significantly and
as an individual moves from the contemplation to action stage, their perceived benefits
decreases significantly.
Using a cross-sectional design, it was found that, among smokers, stage of change
can be predicted by a change in attitudes about the behavior. Specifically, an individual’s
awareness of the burdens of the disorder rises substantially between the precontemplation
and contemplation stages, while the perceived benefits of smoking behavior remain high
until the action stage (Fava, Velicer, & Prochaska, 1995). This same effect has been
found longitudinally, over a period of six months (Velicer et al., 1985). Additionally, a
similar pattern was found when Prochaska and colleagues (1994) examined decisional
balance and stage of change in 12 different behaviors (such as smoking cessation, weight
control, adolescent delinquent behavior, condom use, sunscreen use, and mammography
screening).
Cognitive Representations of the Disorder
The final construct I will be reviewing is perceptions of the disorder. For some
time, health psychologists have expressed that individuals cope with a disorder
differently, based on how they understand it (Leventhal, Nerenz, & Steele, 1984).
Individuals with the exact same disorder may experience different symptoms, may
48
attribute the disorder/symptoms to different causes, and may even perceive similar
symptoms in completely different ways; in other words, each individual has their own
representation of their illness. This representation may closely match a professional
evaluation or may be quite disparate, however there is certainly value in understanding a
patient’s perceptions of their problem.
Leventhal’s Cognitive Representation theory suggests that there are five different
aspects of cognitive representation (Leventhal, Nerenz, & Steele, 1984; Leventhal &
Diefenbach, 1991). These include identity (symptoms associated with the illness),
consequences (perceived effects and outcome), cure control (how one treats the illness),
cause, and timeline (Leventhal, Nerenz, & Steele, 1984). Originally this theory was
developed to assess medical health concerns. Over time, however, the measure has been
adapted to assess mental health concerns as well. In the following sections I will review
how cognitive representations of an individual’s disorder impact their treatment seeking
as well as all relevant illness perception research done with eating disorders.
Cognitive Representations and Treatment Seeking. Three articles were found
that examined cognitive representations and help-seeking (Al Anbar, Dardennes, PradoNetto, Kaye, & Contejean, 2010; Lawson, Bundy, Lyne, & Harvey, 2004; Vanheusden et
al., 2009). All three of these studies reported using the Illness Perception Questionnaire
(IPQ, Weinman, Petrie, Moss-Morris, & Horne, 1996) revised in a way to fit the
particular population sampled (diabetes, general mental health). Use of the IPQ was the
primary commonality in these articles; correlates, methodology, and type of illness
varied. Overall, these studies indicate that there is a significant relation between cognitive
representations of diseases and an individual’s help-seeking behavior.
49
In one study of general mental health concerns, authors found that cognitive
representations of the individual’s mental health problem was related to their mental
health service use (Vanheusden et al., 2009). Specifically, authors sampled 830
individuals who self-identified as having mental health problems over the past year.
Using multivariate logistic regression analyses, the authors reported that, independent of
sex, age, and disorder severity, individuals who perceived their disorder as having more
intra-psychic causes, having greater consequences, and being more responsive to
treatment were more likely to have reported seeking help in the past year. Individuals
who reported having more perceived personal control of the disorder reported
significantly less help-seeking behavior (Vanheusden, 2009).
In a second study, authors examined the relation between cognitive
representations of diabetes and help-seeking (Lawson et al., 2004). The authors sampled
84 diabetic individuals (42 who had not sought care for 18 months and 42 matched
controls who received specialist care). Results indicated that individuals who did not
regularly seek help for their disease were significantly more likely to perceive their
disorder as difficult to control and having more serious consequences than individuals
who did seek regular care (Lawson et al., 2004).
The final study sampled 89 individuals who were parents of a child with autism
spectrum disorder (Al Anbar et al., 2010). The authors assessed cognitive representations
of the child’s illness as well as any help-seeking they had participated in. Researchers
collected information about a range of help-seeking behaviors including medication,
dietary changes, therapy, and educational programs. Results of regression analyses (using
each form of help-seeking behavior as the dependent variable) indicated that there were
50
many significant relations between cognitive representations of the child’s autism
spectrum disorder and particular help-seeking behaviors. Specifically, authors reported
that perceived seriousness of the child’s disorder significantly predicted increased use of
education resources and unpredictable course of the child’s disorder significantly
predicted medication use. Additionally, greater perceived control of the child’s disorder
predicted both reduced use of nutritional treatments and reduced use of pharmaceutical
treatments. Belief that the child’s disorder was primarily caused by hereditary factors and
perception that the child’s disorder was less cylical (more consistent) predicted
attendance at training programs (Al Anbar et al., 2010).
Cognitive Representations and Eating Disorders. Three articles were found
that examined cognitive perceptions of eating disorders (Holliday, Wall, Treasure, &
Weinman, 2005; Marcos, Cantero, Escobar, & Acosta 2007; Stockford, Turner, &
Cooper, 2007). Authors of all three articles reported using a revised version of the IPQ
(Weinman et al., 1996) and reported significant findings; additionally, all researchers and
participants were European (UK and Spain). Aside from these similarities, the three
studies were quite different.
In the first study, authors sought to examine the relation between cognitive
representations of eating disorders and stage of change as well as to develop a reliable
revision of the IPQ for eating disorders (Stockford, Turner, & Cooper, 2007). The authors
sampled 69 individuals, and using the Eating Attitudes Test-40 (EAT-40; Garner &
Garfinkel, 1979) determined diagnosis: one-third with AN, one-third with BN, and onethird with severe EDNOS. Results indicated that there was a significant relation between
stage of change and several IPQ-R subscales.
51
Authors analyzed the data with four hierarchical multiple regression analyses, one
with each stage of change score as the dependent variable. Independent variables
included demographics (age, length of disorder, Body Mass Index), psychological health
measures (Eating Attitudes Test, Beck Depression Inventory, Self-Esteem Scale), social
support, and locus of control were entered first (stage 1), followed by the IPQ-R
subscales. Results indicated that, for the precontemplation stage, the stage 1 predictor
variables accounted for 22% of the variance and adding emotional representation
(emotional distress) and illness coherence (ability to make sense of their symptoms)
scores to the model resulted in a model that explained 45% of the variance; this change
was significant. Regarding the contemplation stage, the stage 1 predictor variables
accounted for 28% of the variance and adding emotional representations to the model
resulted in a model that explained 36% of the variance; this change was significant. Stage
1 predictor variables accounted for 38% of the variance in the action stage and adding the
consequences (awareness of negative consequences of the ED) variable to the model
resulted in a model that explained 45% of the variance; additional variance explained was
significant. Finally, for the maintenance stage, the stage 1 predictor variables accounted
for 25% of the variance; adding treatment control (belief that treatment will improve the
ED) and timeline (belief that disorder is chronic rather than acute) variables, which
resulted in a model that explained 49% of the variance, made a statistically significant
improvement.
In the second study, authors aimed to examine any relation between cognitive
representations of eating disorders and psychosocial adaptation (Marcos et al., 2007). The
authors administered several self-report measures, including a Spanish version of the IPQ
52
revised for eating disorders, to 98 Spanish females with an eating disorder. Mean scores
on IPQ-R subscales (range 5-25) indicated that individuals with an eating disorder
reported they perceived their illness as controllable (M = 20, SD = 2.86), treatable (M =
19, SD = 3.57), and eliciting strong emotion (M = 19, SD = 4.26). Individuals with AN,
BN, and EDNOS reported statistically equal scores on all subscales except for
experiencing a cyclical timeline; on this scale individuals with BN reported a more
noticeable cycle than individuals with AN or EDNOS (t = -2.27, p < .05). Authors also
reported that the emotional representation subscale of the IPQ-R was significantly related
to reported anxiety, depression, and psychological distress. Conversely, insight that many
symptoms experienced (sore throat, stomach pain, sore eyes) are related to the ED was
associated with better emotional and psychosocial adjustment. Individuals who reported a
belief that treatment could be effective reported less anxiety, depression, and illness
distress and greater overall functioning than those who had less faith in treatment
(Marcos et al., 2007).
In the final study, authors sought to compare illness perceptions of Anorexia
Nervosa in individuals with AN and individuals in the general population (Holliday et al.,
2005). Authors recruited 520 individuals with a diagnosed past or present ED, received
data from 195 individuals (with many ED diagnoses) and included in their sample 95
individuals with a current diagnosis of AN, as determined by the Eating Disorder
Examination Questionnaire (EDE-Q, Fairburn & Cooper, 1993) and 80 men and women
with no eating disorder diagnosis. The two groups completed similar IPQ-R scales with
mild wording changes so that individuals in the AN group were asked about their own
experiences and individuals in the layperson group were asked their perceptions of people
53
with Anorexia (Holliday et al., 2005). Individuals with AN reported perceiving many
negative consequences of the disorder (M = 4.08 on a 5-point scale) as well as emotional
distress (M = 4.00 on a 5-pt scale). Results indicated that laypeople had more confidence
than AN individuals that treatment would be effective (M = 3.9 and 3.4 respectively, p <
.001) and laypeople believed the disorder was less distressing than AN individuals (M =
2.6 and 4.0 respectively, p < .001, Holliday et al., 2005).
Rationale for Hypotheses
In the present study I hypothesized that there would be a significant relation
between eating problems and an individual’s attitudes/interest in psychological help
seeking. This hypothesis is consistent with research that demonstrates that individuals
with greater severity are more likely to seek psychological help (Bland et al., 1997; Wang
et al., 2005a). Three measures of help seeking were used to examine differing levels of
help seeking (attitudes, learning about eating disorders and treatment, getting information
about making a counseling appointment). Research demonstrates that individuals may
choose to seek help in one way, but not in another (Al Anbar et al., 2010) and for this
reason, the hypotheses were tested with three different forms of help-seeking.
Furthermore, in the present study I hypothesized that five constructs would
significantly moderate the relation between severity of eating problems and help-seeking
choices: (a) insight about the eating concerns, (b) internalization of cultural beauty
norms, (c) benefits of the eating problems, (d) burdens of the eating problems, and (e)
awareness of the function that the eating problems serve to avoid discomfort. In each of
the five hypotheses, it was suggested that the moderating construct discussed would
significantly change the relation between eating concerns (severity) and help-seeking
54
attitudes/behavior. The rationales for the hypotheses concerning all five constructs are
discussed below.
Insight. No studies were found examining the relation between insight and helpseeking behavior, however there is research examining the relation between insight and
stage of change. It was expected that individuals in the precontemplation stage would be
unlikely to seek help, while those in the contemplation or action stage would be more
open to seeking help. Research indicates that individuals in the precontemplation stage
were more likely to have low insight into eating concerns while individuals in the
contemplation stage were more likely to have higher levels of insight (Stockford, Turner,
& Cooper, 2007). In the present study I hypothesized that severity of ED concerns would
predict willingness to seek help and that insight will moderate this relationship such that
individuals high in insight would be more likely to seek help than those low in insight.
Internalization of cultural beauty norms. No research was found examining the
relation between internalization of cultural beauty norms and help-seeking for an eating
disorder. However, there is clear evidence that high internalization of cultural beauty
norms is positively related with body dissatisfaction and negative eating behaviors
(Forbes, Jobe, & Revak, 2006; Morrison & Sheahan, 2009; Rukavina & Pokrajac-Bulian,
2006).
In the present study I hypothesized that severity of ED concerns would predict
willingness to seek help and that internalization of cultural beauty norms would moderate
this relationship such that individuals who have highly internalized cultural beauty norms
will be less likely to seek help than those with low internalization of cultural beauty
norms. It was expected that individuals who have personally internalized society’s
55
message that one must be thin, attractive, and athletic will perceive their eating attitudes
and behaviors as more consistent with society’s expectations and thus will have less
interest in seeking help than individuals who believe that their behaviors are inconsistent
with society’s messages.
Benefits of eating concerns. While no studies were found examining the relation
between benefits of an eating disorder and help-seeking, research demonstrates that
movement from the contemplation to the action stage of change is marked by a decrease
in reported benefits of the disorder (Prochaska, 1994). It seems reasonable that
individuals moving from the contemplation to action stage of change will be more likely
than individuals in the precontemplation stage to seek help. In the present study I
hypothesized that severity of ED concerns will predict willingness to seek help and that
benefits of the disorder will moderate this relationship such that participants who report a
greater degree of benefits to the disorder will be less likely to seek help than individuals
who report a lesser degree of benefits.
Burdens of eating concerns. No studies were found that studied the relation
between burdens of a disorder and help seeking. However, research demonstrates that
movement from the precontemplation to the contemplation stage of change is marked by
a noted increase in burdens of the disorder (Prochaska, 1994). Moreover, one study
reports that individuals in the contemplation and action stages reported greater burdens of
their eating disorder than individuals in the precontemplation stage (Cockell, Gellar, &
Linden, 2003). It seems reasonable that individuals moving into the contemplation stage
of change or those in the action stage would be more likely to seek help than those
remaining in the precontemplation stage. Thus, in the present study I hypothesized that
56
severity of ED concerns will predict willingness to seek help and that burdens of the
disorder will moderate this relationship such that individuals who report more burdens of
their eating concerns will be more likely to engage in help seeking behavior.
Awareness of functional avoidance. Individuals in the contemplation or action
stage of change report greater awareness of the function that eating concerns serve to
avoid discomfort than individuals in the precontemplation stage (Cockell, Gellar, &
Linden, 2003). It seems reasonable that individuals in the contemplation or action stage
of change are more willing to seek help. Therefore, in the present study I hypothesized
that severity of ED concerns will predict willingness to seek help and that awareness of
functional avoidance will moderate this relationship such that individuals with high
awareness of functional avoidance will be more likely to seek help than individuals with
low awareness of functional avoidance.
57
Chapter 3. Methods
Participants
Data for this study was collected from women at a large Upper Midwestern
University. These students were recruited from introductory psychology courses during
the Fall semester of 2010. Screening and recruitment of participants was done through the
Psychology Department’s Mass Testing. In Mass Testing, several hundred participants
complete demographic information and a multitude of self-report measures at one time. A
total of 1,846 Mass Testing measures were completed over two semesters.
For the present study, participants were screened based on their answers to the
Eating Attitudes Test-26 (EAT-26, Garner et al., 1982). Women who scored in the top
40% of this measure were recruited to participate. Prior research indicates that between
36% and 49% of any given female university sample reports at least moderately
disordered eating (Meyer, 2001, 2005). While both men and women can experience
eating disorders, the disorder primarily affects women. For this reason, only women were
invited to participate in the study. After removing duplicate participants and men, a total
of 1,204 women were screened. A total of 509 women were invited to participate in the
present study. For an analysis using four predictor variables, a sample of 84 would be
required to find a medium sized effect at p < .05 and a sample of 599 would be required
to find a small effect with the same risk of Type I error (Cohen, 1992). Complete data
was collected from 249 women.
Procedures
Students in introductory psychology courses were recruited to participate in the
psychology department’s Mass Testing. They were offered one research credit, which
58
contributes to their class grade; completing all measures was expected to take
approximately 45 minutes. Participants signed up through the University SONA system
and were directed to SurveyMonkey to complete Mass Testing. Participants completed
many self-report measures that were submitted by several researchers. Most of these selfreport measures were not be related to each other. Additionally, participants were asked
to complete demographic information including age, gender, race/ethnicity, and year in
school. As Mass Testing is frequently used to screen participants for future studies, they
were also asked to provide their email for follow-up contact. Before beginning the
survey, participants were provided information about the study and asked to agree to all
informed consent information. After they completed the measures, participants were
given a debriefing form. Any participant under the age of 18 was required to obtain
parental consent before participating in Mass Testing.
For the present study, participants in Mass Testing were administered the Eating
Attitudes Test (EAT-26, Garner et al., 1982) as well as a one-item measure assessing
their interest in seeking therapy for eating concerns (“If you had the resources, would you
be interested in seeking help for eating or body image concerns?”) See Appendix A for
these items. Participants screened into the study (based solely on their gender and EAT26 scores, not their willingness to attend a presentation) were sent an email inviting them
to participate in the present study. This email provided information about the study,
specifically that the researcher is interested in understanding psychological help-seeking
behavior, the format of the measures, the time required, and compensation received for
participating (See Appendix B). The email also invited students to sign-up through the
University SONA system (see SONA posting form in Appendix C), which provided them
59
a link to all measures on SurveyMonkey. The email included a password to sign-up on
SONA so that only screened participants could access the study. Three reminder emails
(See Appendix B) were sent to students who had not yet elected to participate. The first
email was sent one week after the original, the second three weeks after the original, and
the third five weeks after the original. As only a set number of participants could
participate in a SONA experiment each week, reminder emails were primarily aimed at
individuals who wanted to participate, but were unable to in prior weeks due to lack of
availability.
Once students elected to participate in the study, they completed the measure
entirely online, at SurveyMonkey.com, a website that provides survey formatting and
data storage and uses a multitude of security measures to protect collected data.
Participants were given 2 research credits and completed approximately 60-90 minutes
worth of questions. Before beginning the study measures, participants were given an
opportunity for informed consent (see Appendix D); after reading the form, they were
given an option to leave the experiment or to continue. Individuals not yet able to provide
consent because of age were required to provide a signed parent consent form (see
Appendix E). Measures included in this section of the study are as follows: de-labeling
statement and demographic questions (see Appendix F), Clinical Outcomes in Routine
Evaluation – Short Form (CORE-SF, see Appendix G), the Illness Perception
Questionnaire – Revised (IPQ-R, see Appendix H), the Sociocultural Attitudes Towards
Appearance Scale-3 (SATAQ-3, see Appendix I), the Decisional Balance Scale for
Anorexia (DBS, see Appendix J), and the Attitudes Towards Seeking Professional
Psychological Help Scale (ATSPPHS, see Appendix K). The DBS and IPQ-R had minor
60
wording changes, as the past measures were developed with the assumption that
participants had a known, diagnosed eating disorder; participants in this study will likely
not fit those criteria.
After completing these measures, participants were asked two questions, the first
about their interest in learning more about eating concerns (“Would you like to be
directed to online information to learn more about eating issues?”) and the second about
their interest in initiating a therapy appointment (“Would you like to be directed to the
University Student Counseling Service website to sign up for an appointment to talk
about your body image, eating, or exercise? Counseling there is free”, see Appendix L).
These questions are behavioral measures that indicated the degree to which an individual
is open to seeking help for an eating concern. When participants selected “yes” as a
response to either question, an additional web browser window opened, containing the
selected material (See Appendices M & N). After responding to these questions,
participants were directed to the debriefing form (see Appendix O). The University
Internal Review Board has approved both the screening measure in the Mass Testing
study (see Appendix P) as well as the present study (see Appendix Q).
Instruments
All measures given to participants will be discussed here. The screening measure
will be discussed first, followed by the control/covariate measure, moderator variable
measures, and criterion variable measures.
Eating Attitudes Test. The Eating Attitudes Test (EAT-26, Garner et al., 1982) is
a 26 item self-report measure of problems with food and/or body image. The EAT-26 is a
shortened version (using factor analysis) of the original 40-item Eating Attitudes Test
61
(Garner & Garfinkel, 1979). Based on this factor analysis, Garner and colleagues
determined there were three factors to the EAT-26 including “dieting behavior”,
“Bulimia,” and “oral control,” however the authors do no take factor scores into
consideration when determining the implications of an individual’s total score. Sample
items from the EAT-26 include “Am terrified about being overweight,” “Aware of the
calorie content of foods that I eat,” and “Have the impulse to vomit after meals.”
Individuals respond to these items on a six-point Likert scale from “never” to “always.”
Scoring of this measure is unique, in that responders do not start accruing points
on an item until they endorse frequent problems (responses “never,” “rarely,” and
“sometimes” earn 0 points while “often” earns 1 point, “usually” earns 2 points, and
“always” earns 3 points). Scores are calculated by totaling points earned; total scores
range from 0 to 78. Authors suggest using a cutoff of 20 to indicate more significant
eating and body image concerns, but clearly state that this measure does not provide any
diagnosis (Garner & Garfinkel, 1979). In their original article, Garner and Garfinkel
(1979) reported acceptable internal consistency for individuals with Anorexia (α = .79),
and a control group of women with no eating disorder (α = .94). These authors also found
that the original Eating Attitudes Test was able to differentiate between individuals with
Anorexia Nervosa, those who had recovered from AN, and a group of obese females
(Garner & Garfinkel, 1979). Other articles have also reported acceptable consistency
ranging from .75 to .87 (Banasiak, Wertheim, Koerner, & Voudouris, 1999; Morry &
Staska, 2001; Prouty, Protinsky, & Canady, 2002; Rukavina & Pokrajac-Bulian, 2006).
The EAT-26 has also been shown to have good test-retest reliability (r = .89) over a
period of four to five weeks (Banasiak et al., 1999). In this study, the EAT-26 scores
62
range from 7 to 75 and higher scores mean the participant reported more disordered
eating behavior
Clinical Outcomes in Routine Evaluation – Short Form. The Clinical
Outcomes in Routine Evaluation (CORE, Evans et al., 2000) was developed to measure
client distress and improvement throughout therapy. The measure includes 34 items with
four subscales including Well Being, Problems or Symptoms, Functioning, and Risk. The
present study uses three of the four subscales, excluding the six-item Risk subscale.
Sample items include “I have felt optimistic about my future,” “I have felt despairing or
hopeless,” “Unwanted images or memories have been distressing me,” and “I have been
able to do most things I needed to.” Items range from low intensity (“I have felt tense,
anxious, or nervous”) to high intensity (“I have felt panic or terror”). Participants are
asked to respond to items on a 5-point Likert scale ranging from 0 (“not at all”) to 4
(“most or all the time”). The CORE mean score can range from zero to four with higher
scores indicating greater general psychological distress.
Reliability and validity for the CORE-OM are acceptable. Cronbach’s alphas are
reported to be at least .75 for each subscale with a full-scale alpha of .94 (Sinclair,
Barkham, Evans, Connell, & Audin, 2005). Demonstrating concurrent validity, there was
a .86 correlation between the CORE-OM and the Beck Depression Inventory (Evans et
al., 2002). Unlike many proprietary measures, the CORE-OM is ‘copyleft,’ meaning that
it can be used for clinical and research purposes but cannot be changed or used for
financial gain. In the present study, internal consistency was high (α =.94).
Sociocultural Attitudes Towards Appearance Scale-3. The Sociocultural
Attitudes Toward Appearance Scale-3 (SATAQ-3, Thompson, van den Berg, Roehrig,
63
Guarda, & Heinberg, 2004) is a revision of a widely used measure of thin-ideal
awareness and internalization. It measures the degree to which respondents are aware of
appearance messages sent by the media as well as their desire to look like people in the
media. There are four subscales including information (awareness, “TV programs are an
important source of information about fashion and ‘being attractive’”), pressures (“I’ve
felt pressure from TV and magazines to be thin.”), internalization-general (“I would like
my body to look like the people who are on TV”), and internalization-athlete (“I wish I
looked as athletic as people in magazines”). The scale includes 30 items and employs a 5
point Likert scale ranging from 1 (“definitely disagree”) to 5 (“definitely agree”). Scores
are calculated by determining the mean, which can range from 1 to 5, where higher
scores indicate greater internalization of sociocultural beauty norms.
Reported Cronbach’s Alpha scores on these four subscales were consistent and
high, ranging from .89 to .96. Internal consistency in the total scale was also high (α =
.96, .94 in two samples, Thompson et al., 2004). The authors reported convergent validity
estimates with the Ideal Body Internalization scale (correlations with SATAQ-3 subscales
ranged from .37-.53, all at p < .01), with the Eating Disorder Inventory Drive for
Thinness subscale (rs ranged from .17-.58, all at p <.05), and with the Eating Disorder
Inventory Body Dissatisfaction subscale (rs ranged from .17-.51, all at p <.05). The
authors reported that the SATAQ was not significantly correlated with age and that only
one SATAQ-3 (pressures) was correlated with Body Mass Index (rs = .16, .18 in two
samples, p <.05, Thompson et al., 2004). In the present study, internal consistency for the
entire scale was acceptable (α =.942).
64
Illness Perception Questionnaire – Revised. The revised version of the Illness
Perception Questionnaire (IPQ-R, Stockford, Turner, & Cooper, 2007) is an 89-item
measure including 9 subscales intended to measure client perceptions of their own eating
disorder. These subscales include identity, consequences, acute/chronic timeline, personal
control, treatment control/cure, cause, cyclical timeline perceptions, illness coherence,
and emotional representations. Several of these subscales have names that clearly and
accurately describe the items included (consequences, timeline, control, cause), however
others are not as clear. The identity subscale, including 32 items, assesses the degree to
which an individual attributes their symptoms to an eating problem; participants are
asked about a symptom (“avoiding weighing myself”), asked if they have engaged in this
behavior (yes/no), and then asked to rate, on a 5 point Likert scale, “Degree to which I
feel this is related to my food and body image issues.” Scoring of this subscale involves
totaling the Likert scale ratings and dividing by the total symptoms experienced. The
Identity subscale ranges from 0 to 4 and higher scores indicate that they are more likely
to feel each symptoms is related to their food and body image issues. Illness coherence
measures the degree to which the respondent is confused about their symptoms or
disorder and emotional representations measures the degree to which an individual is
distressed by their symptoms.
The IPQ-R used in the present study is modified from the original (Moss-Morris
et al., 2002) in order to fit an eating disorder population. The original measure was shown
to have good internal reliability (α = .79-.89) and test-retest reliability (r = .46-.88; MossMorris et al., 2002). With the current version of the measure, authors found that the cause
subscale had very low internal consistency (α = -.21-.59). This makes sense, however, as
65
client’s attribution of their disorder to one cause likely precludes them from attributing
the cause to another factor. Additionally, the identity subscale had moderate internal
consistency (α = .59). All other subscales had good internal consistency (α = .71-.91,
Stockford, Turner, & Cooper, 2007). Authors reported acceptable test-retest reliability on
all subscales (r = .59-.78) with assessments being taken two weeks apart. Acceptable
internal consistency (α = .90) was found in the present study.
Decisional Balance Scale for Anorexia. The Decisional Balance Scale for
Anorexia (DBS, Cockell, Geller, & Linden, 2002) is a 30-item self-report measure that
was developed using the transtheoretical model of change. In this theory, it is suggested
that an individual’s willingness to change is impacted by two factors: their perceived
benefits and perceived burdens of a particular behavior; the DBS aims to measure the
benefits and burdens of eating concerns. The measure has three factors: benefits (eight
items), burdens (fifteen items), and functional avoidance (seven items). The third factor,
functional avoidance, was unique to Anorexia Nervosa and essentially captured the
function that the disorder serves to help individuals avoid difficult life issues (negative
emotions, growing up, sexuality). Participants respond on a 5-point Likert scale ranging
from “not at all true” to “completely true.” Scores for each item are averaged, thus scale
scores range from 1 to 5, with higher scores indicating more benefits, more burdens, or
more functional avoidance. The authors reported acceptable internal consistency (α = .88
for each subscale) as well as test-retest reliability for the Benefits (r = .71, p < .001),
Burdens (r =.64, p <.001) and Functional Avoidance (r = .70) subscales over an 8 day
period (Cockell, Geller, Linden, 1999).
66
In the present study, some wording changes were made. The measure was
developed to capture benefits and burdens of Anorexia and many participants in the
present study will not fit this diagnosis. The word “Anorexia” in the original measure was
replaced with “issues with food and body image” in the present study. For example,
“Because of my issues with Anorexia, I feel guilty a lot of the time” was changed to
“Because of my issues with food and body image, I feel guilty a lot of the time.” Other
sample items include “It bothers me that my weight controls my mood,” “I worry about
the effect my issues with food and body image are having on my health,” and “My issues
with food and body image protect me from the difficulties of adult life” (Cockell, Geller,
& Linden, 2002). Internal consistency in the current sample was acceptable for the
Benefits (α = .88), Burdens (α = .93), and Functional Avoidance (α =.86) subscales.
Attitudes Toward Seeking Professional Psychological Help Scale. Attitudes
about seeking therapy will be measured with the Attitudes Toward Seeking Professional
Psychological Help Scale (ATSPPHS, Fischer & Farina, 1995). This measures has been
shortened from the original 29-item scale (ATSPPHS, Fischer & Turner, 1970) to the
current 10 item scale. Sample items include “The idea of talking about problems with a
psychologist strikes me as a poor way to get rid of emotional conflicts” and “I might want
to have psychological counseling in the future” (reverse coded). Participants respond to
items with a 4-point Likert scale ranging from 1 (disagree) to 4 (agree). Scores can range
from 1 to 4 and higher scores indicate greater stigma towards counseling. Research
indicates that the original measure and shortened measure are highly correlated (r = .87,
Fischer & Farina, 1995). Authors also reported good internal consistency (α = .84) and
67
acceptable one-month test-retest reliability (r = .80, Fischer & Farina, 1995). In the
present study, internal consistency was acceptable (α = .81).
Behavioral Measures. Two behavioral measures were used to assess a
participants’ willingness to seek help. The first is about an interest in learning more about
eating concerns and their treatment (“Would you like to be directed to online information
to learn more about eating issues?”) and the second is about their interest in initiating a
therapy appointment (“Would you like to be directed to the University Student
Counseling Service website to sign up for an appointment to talk about your body image,
eating, or exercise? Counseling there is free”, see Appendix M) Students who responded
“yes” to the item about learning more about eating issues were directed to the Something
Fishy webpage, designed to gently inform individuals and their families about negative
eating behaviors and attitudes. See Appendix N for an image of the homepage. Students
who responded “yes” to the item about signing up for a counseling appointment were
directed to the University Student Counseling Service webpage. See Appendix O for an
image of the homepage. Scoring was simply whether the participant responded “yes” or
“no” to each of these items.
Hypotheses
All hypotheses examined the mediating effects of varying constructs on the relation
between eating problems, as measured by the Eating Attitudes Test-26, and three separate
measures of help seeking: (a) Attitudes Towards Seeking Professional Psychological
Help Scale, (b) whether or not an individual chooses to click on a link to gather more
information about eating issues and their treatment and (c) whether or not an individual
chooses to click on a link to learn how to set up a counseling appointment at the
68
University Counseling Service. Each hypothesis below was run with all three measures of
help seeking.
1. After controlling for general distress, measured by the CORE-SF, it was
hypothesized that insight about eating concerns, as measured by the IPQ-R
Identity subscale, wouldl moderate the relation between eating problems and
psychological help seeking, operationalized as the three measures discussed
above, such that for all three measures of help seeking there would be a
significant relation between eating concerns and psychological help seeking for
individuals with a high degree of insight and no significant relation for individuals
with a low degree of insight.
2. After controlling for general distress, measured by the CORE-SF, it was
hypothesized that internalization of societal beauty norms, as measured by the
SATAQ, will moderate the relation between eating problems and psychological
help seeking, operationalized as the three measures discussed above, such that for
all three measures of help seeking there would be a significant relation between
eating concerns and psychological help seeking for individuals with a low degree
of internalization of societal beauty norms and no significant relation for
individuals with a high degree of insight.
3. After controlling for general distress, measured by the CORE-SF, it was
hypothesized that benefits of the eating concerns, as measured by the DBS
Benefits subscale, will moderate the relation between eating problems and
psychological help seeking, operationalized as the three measures discussed
above, such that for all three measures of help seeking there would be a
69
significant relation between eating concerns and psychological help seeking for
individuals with a low level of benefits and no significant relation for individuals
with a high level of benefits.
4. After controlling for general distress, measured by the CORE-SF, it was
hypothesized that burdens of the eating concerns, as measured by the DBSBurdens subscale, will moderate the relation between eating problems and
psychological help seeking, operationalized as the three measures discussed
above, such that for all three measures of help seeking there would be a
significant relation between eating concerns and psychological help seeking for
individuals with a high level of burdens and no significant relation for individuals
with a low level of burdens.
5. After controlling for general distress, measured by the CORE-SF, it was
hypothesized that awareness of the function that eating issues serve to avoid
discomfort, as measured by the DBS Functional Avoidance subscale, will
moderate the relation between eating problems and psychological help seeking,
operationalized as the three measures discussed above, such that for all three
measures of help seeking there would be a significant relation between eating
concerns and psychological help seeking for individuals with a high degree of
awareness of functional avoidance and no significant relation for individuals with
a low degree of awareness of functional avoidance.
70
Chapter 4. Results
In this section, I will discuss the results. First, I will talk about the descriptive
statistics and preliminary analyses performed to better understand the data. Next, I will
report the analyses for each of the hypotheses. Finally, I will discuss additional analyses
done to better understand the results.
Descriptive Statistics and Preliminary Analyses
This sample consisted of 249 undergraduate college women. Means and standard
deviations for demographic (age) and the eight measured variables (EAT score, SATAQ,
ATSPPHS, CORE, IPQ-R Identity, DBS Benefits, DBS Burdens, and DBS Functional
Avoidance) are reported in Table 1. Additionally, to best understand the sample and to
compare it with previous samples in the literature, a variable indicating clinical severity
was calculated. To do this, two scales were used including the Eating Attitudes Test
(EAT-26, Garner, Olmsted, Bohr, & Garfinkel, 1982) and the Eating Disorder Diagnostic
Scale (EDDS, Stice, Telch, & Rizvi, 2000). Participants who met the clinical cutoff for
either of these measures (a sum score of 20 or above on the EAT-26 or symptoms
consistent with a diagnosis of Anorexia Nervosa or Bulimia Nervosa on the EDDS) were
categorized as having a clinical level of severity. There were 48 out of the 249
participants in this sample that were categorized as having a clinical level of severity.
Table 2 provides means and standard deviations for individuals both above and below
this clinical level. Finally, zero-order correlations for the sample are compiled in Table 4.
Because so many correlations were calculated, only correlations significant at the p < .01
level will be discussed.
71
In order to best understand this sample, it is important to compare these
participants to participants in the literature. I’ve primarily done this by comparing mean
scores. First, I examined general psychological distress. In the study where the measure
was developed, authors sampled 483 college women and reported a mean score of 1.07
(SD = .71) in the non-clinical sample and a mean score of 2.16 (SD = .66) in the clinical
sample (Sinclair, et al., 2005). The overall mean score in the present study is 2.30 (SD =
.68). Additionally, individuals above the clinical disordered eating cutoff reported
significantly higher general psychological distress (mean = 2.62) than those below the
cutoff (mean = 2.23). Overall, it appears that this sample reports more distress than a nonclinical, college sample and a comparable amount of distress to a sample of clinical-level
college women. As the present sample screened out individuals with the lowest 60% of
scores on the Eating Attitudes Test, it seems reasonable that the present sample is more
similar to the clinical than non-clinical college student groups.
Authors of the IPQ-R identity scale (measuring insight) did not report an overall
mean score for this scale, but did report that all of the IPQ-R subscales (including
identity) ranged from 3.13-3.96 (SD = .69-.80) on a 5 point scale (Stockford, Turner, &
Cooper, 2007). The present sample mean was 2.68 (SD = .84) on a 5 point scale.
Additionally, the authors reported that participants endorsed experiencing 24 of the 32
symptoms (SD = 4.65), while participants in this study endorsed experiencing
approximately 15 (SD = 5.23). The differences between the present sample and the
previous sample are likely due to severity; in the present sample approximately 20% of
the sample would meet clinical criteria for an eating disorder diagnosis, while in the
previous sample, all participants had a diagnosed eating disorder (Stockford, Turner, &
72
Cooper, 2007). There was a significant difference in the present sample between
individuals who were above and below the clinical cutoff such that individuals who likely
meet criteria for an eating disorder reported greater insight (mean = 3.00) and more
symptoms (18.13) than those who did not meet criteria (mean = 2.61, symptoms = 14).
In the present study, participants reported a mean score on the Sociocultural
Attitudes Toward Appearance Questionnaire (SATAQ, Heinberg, Thompson, & Stormer,
1995) of 3.61 (SD = .64) on a 5 point scale. This can be compared to several samples in
previous studies. First, Mitchell and Mazzeo reported a mean score of 3.18 on the 14item internalization subscale of this measure for a sample of 408 European American
female University students (2009). In another study, authors sampled women ages 13-40
and reported means for both eating disordered participants (Anorexia = 3.59, Bulimia =
3.93, and EDNOS = 3.92) and a control group with no diagnosis (3.01) using the 14-item
internalization subscale (Griffiths, et al., 1999). Finally, authors of a third study sampled
a group of 140 Univeristy women and reported a mean score of 3.32 on the
internalization subscale (Morrison & Sheahan, 2009). Based on this information, it
appears that participants in the present study report a level of internalized societal beauty
norms between those of eating disordered and non-eating disordered women. There was
no significant difference between mean scores for individuals who were above and below
the clinical cutoff.
Students in the present study reported a mean score of 2.35 (SD = .53) on the
Attitudes Toward Seeking Professional Psychological Help Scale (ATSPPHS, Fischer &
Farina, 1995). There was no significant difference between the mean scores for
individuals above and below the clinical cutoff. No previous studies exist using this
73
measure with a sample of individuals with disordered eating concerns. In a study
designed to examine the effects of social network on attitudes toward counseling, authors
sampled 746 undergraduate college students and reported a mean of 2.50 on a 10-item
version of the ATSPPHS (Vogel, et al., 2007). The present sample and the past sample
seem to have similar attitudes towards counseling.
Means for all three of the Decisional Balance Scale for Eating Disorders (Cockell,
Geller, & Linden, 2002) subscales were examined. Authors of the previous study
reported means for each subscale as related to each of Prochaska and DiClimente’s stage
of change. All participants in the previous study met criteria for a current or subthreshold
diagnosis of Anorexia Nervosa and were recruited at a first appointment at an eating
disorder clinic. In the present study, the overall mean for reported benefits of the disorder
was 2.79 (SD = .92) on a 5 point scale; there was a significant mean difference for
individuals above the clinical cutoff (mean = 3.39) and below the cutoff (mean = 2.65).
In the previous study, authors reported means for each stage of change including
Precontemplation (mean = 3.41, SD = .87), Contemplation (mean = 3.17, SD = .95), and
Action (mean = 3.15, SD = .98 Cockell, Geller, & Linden, 2003). In the present study, it
is clear that individuals who report greater disordered eating behavior also experience
more benefits of the disorder than those below clinical threshold and report a similar
degree of benefits as individuals in the Precontemplation stage in the previous study.
Individuals also reported perceived burdens, or drawbacks, of the disorder. In the
present study, the overall mean score was 2.20 (SD = .87) on a 5 point scale; there was a
significant mean difference for individuals above the clinical cutoff (mean = 2.76) and
below the cutoff (mean = 2.07). In the previous study, authors reported means for each
74
stage of change including Precontemplation (mean = 3.34, SD = .70), Contemplation
(mean = 3.83, SD = .55), and Action (mean = 3.91, SD = .87, Cockell, Geller, & Linden,
2003). Scores in the present study appear quite a bit lower than those from the previous
study. This may be due to the severity of disorder experienced by participants in each
sample and/or the fact that all participants in the previous study were seeking counseling.
The final subscale of the Decisional Balance Scale is Functional Avoidance; it
measures the degree to which participants are aware that their disorder helps them to
avoid uncomfortable emotions. In the present study, the overall mean score was 1.60 (SD
= .75) on a 5 point scale; there was a significant mean difference for individuals above
the clinical cutoff (mean = 2.08) and below the cutoff (mean = 1.49). In the previous
study, authors reported means for each stage of change including Precontemplation
(mean = 2.25, SD = .90), Contemplation (mean = 3.01, SD = 1.05), and Action (mean =
3.31, SD = 1.04, Cockell, Geller, & Linden, 2003). It appears that mean scores in the
present study are lower than those from the previous study; again, this may be due to the
severity of disorder experienced by participants in each sample and/or their choice to
seek counseling. Individuals above the clinical cutoff in the present study have scores
similar to individuals in the Precontemplation stage in the previous study.
Because the Eating Attitudes Test (EAT-26, Garner, Olmsted, Bohr, & Garfinkel,
1982) was used to screen out participants, I will describe this sample’s scores but will not
compare the present mean to other samples. The mean score for the present sample is
18.1, where a score above 20 likely indicates a clinical eating disorder diagnosis.
Participants above the clinical cutoff had a mean EAT-26 score of 36.4 and those below
the clinical cutoff had a mean score of 14.6.
75
Tables 4 and 5 present the means and standard deviations of the variables of
interest in the study by the two behavioral measures. Yes and no responses from the two
behavioral measures were considered as independent variables. The variables of interest
served as the dependent variables. Two sets of five Analyses of Variance (ANOVA) were
conducted with the SATAQ, ATSPPHS, CORE, IPQ-R, and the EAT scales. Two
Multivariate Analyses of Variance (MANOVA) were conducted with the three DBS
scales entered simultaneously. The alpha was adjusted to .01 to control for multiple tests.
As can be seen in Table 3, there were significant mean differences between
women who did and did not want to view online information about disordered eating on
three scales. Women who reported a desire to view online information about disordered
eating, compared to women who did not desire to view this information, reported
significantly higher scores on the CORE (F[1, 248] = 8.54, p = .004). The MANOVA
examining all three DBS subscales was also significant (Wilkes Λ = .93, F[3, 245] =
5.73, p = .001). The follow up univariate ANOVAs showed that the DBS-Burdens
subscale (F[1, 248] = 29.84, p < .001), and the DBS-Functional Avoidance subscale
(F[1, 248] = 17.78, p < .001) were both significant at the p < .001. That is, women who
indicated a desire to view online information about eating issues, when compared with
women who did not indicate a desire to view this information, reported more general
psychological distress, more drawbacks to their disordered eating behavior, and a greater
awareness of the function their eating disorder serves to avoid uncomfortable emotion.
There were also significant mean differences between women who did and did not
want to schedule a counseling appointment, as can be seen in Table 4. Women who
indicated a desire to schedule a counseling appointment, compared to those who did not
76
want to schedule an appointment, reported significantly lower scores on the ATSPPHS
(F[1, 248] = 6.67, p = .010). The MANOVA examining all three DBS subscales was also
significant (Wilkes Λ = .89, F[3, 245] = 10.28, p < .001). The follow up univariate
ANOVAs showed that the DBS-Burdens subscale (F[1, 248] = 13.16, p < .001) and the
DBS-Functional Avoidance subscale (F[1, 248] = 8.50, p = .004) were both statistically
significant. That is, women who indicated a desire to schedule a counseling appointment,
when compared with women who indicated no desire in scheduling an appointment,
indicated a more positive attitude toward counseling (lower stigma toward counseling),
more drawbacks to their disordered eating behavior, and a greater awareness of the
function their eating behaviors serve to avoid uncomfortable emotion.
Additionally, correlations with at least a moderate effect size (r > .25) and
significant at the p < .01 level were examined as can be seen by Table 4. Three of the
scales discussed are subscales of a larger measure (DBS-Benefits, DBS-Burdens, and
DBS-Functional Avoidance). They will be discussed separately, as the literature indicates
each construct has a different connection with the change process; however it should be
expected that they will be highly correlated. Age and the ATSPPHS were not
significantly correlated with any other variables.
As expected, several of the largest correlations existed between the DBS
subscales. There were significant positive correlations between DBS-Benefits and DBSBurdens scales (r = .58), DBS-Benefits and DBS-Functional Avoidance scales (r = .44),
and DBS-Burdens and DBS-Functional Avoidance scales (r = .70). A large correlation
also existed between the IPQ-R identity scale and the DBS-Benefits scale (r = .61) as
well as the DBS-Burdens scale (r = .57). These correlations indicate that as individuals
77
reported a greater total of any of these three scales, scores on the other scales were higher
as well.
Additionally, there were significant medium-sized correlations between the
CORE and DBS-Burdens scale (r = .47) as well as the DBS-Functional Avoidance scale
(r = .42), indicating that individuals who reported greater general psychological distress
also reported more drawbacks to their eating concerns as well as more awareness of the
function their eating behaviors serve to avoid uncomfortable emotion. Correlations of a
similar magnitude also existed between the SATAQ and IPQ-R identity scale (r = .43)
and DBS-Benefits scale (r = .41), suggesting that individuals who reported a greater
internalization of societal beauty norms also reported a greater insight that each of their
individual symptoms was related to one underlying cause as well as greater benefits to
their eating behaviors.
Medium sized correlations were also found between the Eating Attitudes Test and
several other scales including the IPQ-R identity (r = .35), DBS-Benefits (r = .33), DBSBurdens (r = .35), DBS-Functional Avoidance (r = .33), and CORE scales (r = .26),
meaning that women who reported more disordered eating behavior also reported more
insight that each of their symptoms was related to an underlying cause, more benefits to
their eating behaviors, more drawbacks to their eating behaviors, a greater awareness of
the function their disorder serves to avoid uncomfortable emotion, and greater general
psychological distress. There were also a medium sized correlations between the DBSFunctional Avoidance scale and the IPQ-R identity scale (r = .39), indicating that
participants who reported a greater awareness of the function their eating behaviors serve
to avoid uncomfortable emotion also reported more insight that each of their individual
78
symptoms is related to one underlying disorder. Additionally, a small to medium
correlation existed between the CORE and DBS-Benefits scale (r = .32) and IPQ-R
identity (r = .23), indicating that women who reported greater general psychological
distress also reported more benefits to their disorder and greater insight that each of their
eating symptoms are related to one underlying disorder.Finally, a small correlation was
found between the SATAQ and DBS-Burdens scale (r = .29), suggesting that individuals
who reported more internalized societal beauty norms also reported more drawbacks to
their eating behaviors.
Frequencies of yes/no responses to both of the categorical criterion variables were
examined. Out of 249 participants, 30 (12%) indicated they would like to be directed to
online information about eating issues and 12 (2.8%) indicated they would like to be
directed to the University Counseling Service website to make an appointment to discuss
eating or body image concerns. All women who indicated a desire to make a counseling
appointment also marked that they wanted to be directed to online information about
eating issues.
Hypotheses
The Statistical Package for the Social Sciences (SPSS 16, 2008) was used for all
analyses. Additionally, for all analyses, I created an interaction term by multiplying the
hypothesized standardized moderator variable and standardized predictor variable. To test
all hypotheses predicting attitudes toward help-seeking, hierarchical moderated
regression analyses were used (Wampold & Freund, 1987). For each of the five
hypotheses, variables were entered in the following order: (1) the CORE, (2) the Eating
79
Attitudes Test and hypothesized predictor, and (3) the interaction term. See Table 6 for
specific constructs.
To test all hypotheses predicting group membership of whether or not a
participant chose to seek more information or help, binary logistic regressions were used
(Peng, Lee, & Ingersol, 2002). For each of the five moderators, variables were entered in
three models: (1) the CORE, (2) the CORE, the predictor variable, and the hypothesized
moderator, and (3) the CORE, Eating Attitudes Test, hypothesized moderator, and the
interaction term. See Table 6 for specific constructs.
Hypothesis 1. For both the categorical and continuous dependent variables, it was
hypothesized that, after entering the CORE as a control variable, the IPQ-R identity
measure would moderate the relation between the EAT-26 and psychological help
seeking such that there would be a significant relation between the EAT-26 and
psychological help seeking for individuals with higher IPQ-R identity scores and no
significant relation for individuals with lower IPQ-R identity scores.
A hierarchical multiple regression was conducted to test this hypothesis for the
ATSPPHS. As shown in Table 7, Step 1 was not significant (F [1, 248] = .07, p = .79)
and predicted less than one percent of the variance. Additionally, adding the EAT-26
score and the IPQ-R identity score in Step 2 did not lead to a significant effect (F [3, 248]
= .26, p = .86). Finally, in Step 3 the interaction between the EAT-26 and IPQ-R identity
was added to the model; this analysis was again not significant (F [4, 248] = .43, p = .79)
and predicted less than one percent of the variance.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual reported a desire to be directed to online information about
80
eating issues. As shown in Table 8, Step 1 was significant (χ 2 [1, 248] = .7.96, p = .005)
and predicted less than 3 percent of the variance using the Cox and Snell (1989) R Square
(R2) statistic. Additionally, adding the EAT-26 score and the IPQ-R identity score in Step
2 did not lead to a significant improvement (χ 2 [2, 246] = .1.40, p = .497) although the
overall model remained significant (χ 2 [3, 245] = 9.36; p = .025). Finally, in Step 3 the
interaction between the EAT-26 and IPQ-R identity was added to the model; this analysis
did not lead to a significant improvement in the overall model (χ 2 [1, 247] = .156, p =
.693) although the model remained significant (χ 2 [4, 244] = 9.51, p = .05) and predicted
3.7 percent of the variance. Table 9 shows the standardized coefficients, the odds ratios,
and the confidence intervals for the odds ratio. As can be seen by Table 9, only the
CORE was a significant predictor in the binary logistic regression.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual indicated a desire to schedule a counseling appointment. As
shown in Table 8, Step 1 was not significant (χ 2 [1, 248] = 2.72, p = .099) and predicted
approximately one percent of the variance using the Cox and Snell R Square (1989) (R2)
statistic. Additionally, adding the EAT-26 score and the IPQ-R identity score in Step 2
did not lead to a significant improvement (χ 2 [2, 246] = 1.12; p = .572) and the overall
model remained non-significant (χ 2 [3, 245] = 3.83; p = .28). Finally, in Step 3 the
interaction between the EAT-26 and IPQ-R identity was added to the model; this analysis
did not lead to a significant improvement in the overall model (χ 2 [1, 247] = .107, p =
.74) and, again, the overall model was not significant (χ 2 [4, 244] = 3.94, p = .414) and
predicted 1.6 percent of the variance. Table 10 shows the standardized coefficients, the
odds ratios, and the confidence intervals for the odds ratio.
81
Hypothesis 2. It was hypothesized that, after entering the CORE as a control
variable, the SATAQ would moderate the relation between the EAT-26 and
psychological help seeking such that there would be a significant relation between the
EAT-26 and psychological help seeking for individuals with lower SATAQ scores and no
significant relation for individuals with a higher SATAQ scores.
A hierarchical multiple regression was used to test this hypothesis for the
ATSPPHS. As shown in Table 11, Step 1 was not significant (F [1, 248] = .07, p = .79)
and predicted virtually none of the variance. Additionally, adding the EAT-26 and
SATAQ scores in Step 2 did not produce a significant model (F [3, 248] = .12, p = .95).
Finally, in Step 3, the interaction of the EAT-26 and SATAQ was added to the model;
this analysis was again not significant (F [4, 248] = .13, p = .97) and predicted only 0.2%
of the variance.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual reported a desire to view online information about eating
issues. As shown in Table 12, Step 1 was significant (χ 2 [1, 248] = .7.96, p = .005) and
predicted 3 percent of the variance using the Cox and Snell (1989) R Square (R2)
statistic.Additionally, adding the EAT-26 score and the SATAQ score in Step 2 did not
lead to a significant improvement (χ 2 [2, 246] = .36, p = .836) although the overall
model remained significant (χ 2 [3, 245] = 8.32, p = .04). Finally, in Step 3 the interaction
between the EAT-26 and the SATAQ was added to the model; this analysis did not lead
to a significant improvement in the overall model (χ 2 [1, 247] = .05, p = .819) and the
overall model was not significant (χ 2 [4, 244] = 8.37, p = .08) and predicted 3.3 percent
of the variance. Table 13 shows the standardized coefficients, the odds ratios, and the
82
confidence intervals for the odds ratio. As can be seen by Table 13, only the CORE was a
significant predictor in the binary logistic regression.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual indicated a desire to schedule a counseling appointment. As
shown in Table 12, Step 1 was not significant (χ 2 [1, 248] = 2.72, p = .099) and
predicted approximately one percent of the variance using the Cox and Snell (1989) R
Square (R2) statistic . Additionally, adding the EAT-26 score and the SATAQ score in
Step 2 did not lead to a significant improvement (χ 2 [2, 246] = .49, p = .781) and the
overall model remained non-significant (χ 2 [3, 245] = 3.21; p = .36). Finally, in Step 3
the interaction between the EAT-26 and the SATAQ was added to the model; this
analysis did not lead to a significant improvement in the overall model (χ 2 [1, 247] =
1.69, p = .19) and, again, the overall model was not significant (χ 2 [4, 244] = 4.90, p =
.298) and predicted 1.9 percent of the variance. Table 14 shows the standardized
coefficients, the odds ratios, and the confidence intervals for the odds ratio.
Hypothesis 3. It was hypothesized that, after entering the CORE as a control
variable, the DBS-Benefits scale would moderate the relation between the EAT-26 and
psychological help seeking such that there would be a significant relation between the
EAT-26 and psychological help seeking for individuals with low DBS-Benefits score and
no significant relation for individuals with high DBS-Benefits score.
A hierarchical regression was conducted to test this hypothesis for the ATSPPHS.
As shown in Table 15, Step 1 was not significant (F [1, 248] = .07, p = .79) and predicted
less than one percent of the variance. Additionally, adding the EAT-26 and DBS-Benefits
scores in Step 2 did not produce a significant model (F [3, 248] = .58, p = .63). Finally,
83
the interaction of the EAT-26 and DBS-Benefits was added in Step 3; this model was
again not significant (F [4, 248] = .54, p = .71) and predicted nearly one percent of the
variance.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual reported a desire to be directed online information about
eating issues. As shown in Table 16, Step 1 was significant (χ 2 [1, 248] = .7.96, p =
.005) and predicted 3 percent of the variance using the Cox and Snell (1989) R Square
(R2) statistic . Additionally, adding the EAT-26 score and the DBS-Benefits score in Step
2 did not lead to a significant improvement (χ 2 [2, 246] = 3.06, p = .216) although the
overall model remained significant (χ 2 [3, 245] = 11.02, p = .012). Finally, in Step 3 the
interaction between the EAT-26 and the DBS-Benefits was added to the model; this
analysis did not lead to a significant improvement in the overall model (χ 2 [1, 247] =
.2.80, p = .09), however the overall model was significant (χ 2 [4, 244] = 13.82, p = .008)
and predicted 5.4 percent of the variance. Table 17 shows the standardized coefficients,
the odds ratios, and the confidence intervals for the odds ratio. As can be seen by Table
17, only the CORE was a significant predictor in the binary logistic regression.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual indicated a desire to schedule a counseling appointment. As
shown in Table 16, Step 1 was not significant (χ 2 [1, 248] = 2.72, p = .099) and
predicted approximately one percent of the variance using the Cox and Snell (1989) R
Square (R2) statistic . Additionally, adding the EAT-26 score and the DBS-Benefits score
in Step 2 did not lead to a significant improvement (χ 2 [2, 246] = .35, p = .838) and the
overall model remained non-significant (χ 2 [3, 245] = 3.07; p = .381). Finally, in Step 3
84
the interaction between the EAT-26 and the DBS-Benefits was added to the model; this
analysis did not lead to a significant improvement in the overall model (χ 2 [1, 247] =
1.07, p = .301) and, again, the overall model was not significant (χ 2 [4, 244] = 4.14, p =
.388) and predicted 1.6 percent of the variance. Table 18 shows the standardized
coefficients, the odds ratios, and the confidence intervals for the odds ratio.
Hypothesis 4. It was also hypothesized that, after entering the CORE as a control
variable, the DBS-Burdens scale would moderate the relation between the EAT-26 and
psychological help seeking such that there would be a significant relation between the
EAT-26 and psychological help seeking for individuals with a high DBS-Burdens score
and no significant relation for individuals with a low DBS-Burdens score.
A hierarchical regression was conducted to test this hypothesis for the ATSPPHS.
As shown in Table 19, Step 1 was not significant (F [1, 248] = .07, p = .79) and predicted
less than one percent of the variance. However, adding the EAT-26 and DBS-Burdens
scores in Step 2 did produce a significant model (F [3, 248] = 2.72, p < .05, ΔR2 = .032)
that predicted approximately 3% of the variance. Finally, the interaction of the EAT-26
and DBS-Burdens scale was added to the model; this analysis was again significant (F [4,
248] = 2.55, p <.05, ΔR2 = .01) and predicted 4% of the variance. As can be seen in
Table 20, only DBS-Burdens significantly contributed to the complete model.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual reported a desire to view online information about eating
issues. As shown in Table 21, Step 1 was significant (χ 2 [1, 248] = .7.96, p = .005) and
predicted 3 percent of the variance using the Cox and Snell (1989) R Square (R 2) statistic.
Adding the EAT-26 score and the DBS-Burdens score in Step 2 lead to a significant
85
improvement (χ 2 [2, 246] = 20.59, p <.001) and the overall model remained significant
(χ 2 [3, 245] = 28.54, p < .001). Finally, in Step 3 the interaction between the EAT-26
and the DBS-Benefits was added to the model; this analysis did not lead to a significant
improvement in the overall model (χ 2 [1, 247] = .11, p = .744), however the overall
model remained significant (χ 2 [4, 244] = 28.65, p < .001) and predicted 10.9 percent of
the variance. Table 22 shows the standardized coefficients, the odds ratios, and the
confidence intervals for the odds ratio. As can be seen by Table 22, only the DBSBurdens scale was a significant predictor in the binary logistic regression.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual indicated a desire to schedule a counseling appointment. As
shown in Table21, Step 1 was not significant (χ 2 [1, 248] = 2.72, p = .099) and predicted
approximately one percent of the variance using the Cox and Snell (1989) R Square (R 2)
statistic . Adding the EAT-26 score and the DBS-Burdens score in Step 2 lead to a
significant improvement in the model (χ 2 [2, 246] = 9.34, p = .009) and the overall
model remained significant (χ 2 [3, 245] = 12.06; p = .007). Finally, in Step 3 the
interaction between the EAT-26 and the DBS-Burdens was added to the model; this
analysis did not lead to a significant improvement in the overall model (χ 2 [1, 247] =
1.21, p = .27) and, again, the overall model was significant (χ 2 [4, 244] = 13.26, p = .01)
and predicted 5.2 percent of the variance. Table 23 shows the standardized coefficients,
the odds ratios, and the confidence intervals for the odds ratio. As can be seen by Table
23, only the DBS-Burdens scale was a significant predictor in the binary logistic
regression.
86
Hypothesis 5. Finally, it was hypothesized that, after entering the CORE as a
control variable, the DBS-Functional Avoidance scale would moderate the relation
between the EAT-26 and psychological help seeking such that there would be a
significant relation between the EAT-26 and psychological help seeking for individuals
with a high DBS-Functional Avoidance score and no significant relation for individuals
with a low DBS-Functional Avoidance score.
A hierarchical regression was used to test this hypothesis for the ATSPPHS. As
shown in Table 24, Step 1 was not significant (F [1, 248] = .07, p = .79) and predicted
less than one percent of the variance. Additionally, adding the EAT-26 and DBSFunctional Avoidance scores in Step 2 did not produce a significant model (F [3, 248] =
.248, p = .86). Finally, in Step3 the interaction between the EAT-26 and DBS-Functional
Avoidance was added to the model; this analysis was again not significant (F [4, 248] =
.19, p = .95) and predicted less than one percent of the variance.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual reported a desire to see online information about eating
issues. As shown in Table 25, Step 1 was significant (χ 2 [1, 248] = .7.96, p = .005) and
predicted 3 percent of the variance using the Cox and Snell (1989) R Square (R 2) statistic
. Additionally, adding the EAT-26 score and the DBS-Functional Avoidance score in
Step 2 lead to a significant improvement in the model (χ 2 [2, 246] = 8.91, p = .012) and
the overall model remained significant (χ 2 [3, 245] = 16.87, p = .001). Finally, in Step 3
the interaction between the EAT-26 and the DBS-Functional Avoidance was added to the
model; this analysis did not lead to a significant improvement in the overall model (χ 2 [1,
247] = 1.87, p = .17), however the overall model was significant (χ 2 [4, 244] = 18.74, p
87
= .001) and predicted 7.2 percent of the variance. Table 26 shows the standardized
coefficients, the odds ratios, and the confidence intervals for the odds ratio. As can be
seen by Table 26, only the DBS-Functional Avoidance scale was a significant predictor
in the binary logistic regression.
A binary logistic regression was conducted to test this hypothesis regarding
whether or not an individual indicated a desire to schedule a counseling appointment. As
shown in Table 25, Step 1 was not significant (χ 2 [1, 248] = 2.72, p = .099) and
predicted approximately one percent of the variance using the Cox and Snell (1989) R
Square (R2) statistic . Additionally, adding the EAT-26 score and the DBS-Functional
Avoidance score in Step 2 did not lead to a significant improvement (χ 2 [2, 246] = 4.15,
p = .13) and the overall model remained non-significant (χ 2 [3, 245] = 6.86; p = .076).
Finally, in Step 3 the interaction between the EAT-26 and the DBS-Functional
Avoidance was added to the model; this analysis did not lead to a significant
improvement in the overall model (χ 2 [1, 247] = 1.15, p = .284) and, again, the overall
model was not significant (χ 2 [4, 244] = 8.01, p = .091) and predicted 3.2 percent of the
variance. Table 27shows the standardized coefficients, the odds ratios, and the confidence
intervals for the odds ratio. As can be seen by Table 27, only the DBS-Functional
Avoidance scale was a significant predictor in the binary logistic regression.
Additional Analyses
In order to better understand these results, additional analyses were conducted.
Most of these involve the construct I created to indicate whether or not the participant
was likely to meet diagnostic criteria for an eating disorder. As discussed above, if a
participant met clinical criteria either on the EAT-26 or the EDDS, they were identified
88
as clinical (48 participants); those who did not meet this criteria were identified as not
clinical (201 participants). See Table 2 for means and standard deviations by clinical
category.
There were significant mean differences between the clinical and non-clinical
groups on six different scales (three of which are subscales of one larger scale, see Table
2). Individuals in the clinical group reported significantly higher scores on the CORE
(F[1, 247] = 13.89, p < .001) as well as IPQ-R identity (F[1, 247] = 8.69, p = .004),
suggesting that women with a clinical level of eating concerns reported more general
psychological distress as well as more insight that each of their symptoms is related to a
single underlying disorder than women below the clinical level. Also, individuals in the
clinical group reported significantly higher scores on the DBS-Benefits scale (F[1, 247] =
21.40, p < .001), the DBS-Burdens scale (F[1, 247] = 18.37, p < .001), and the DBSFunctional Avoidance scale (F[1, 247] = 13.45, p < .001), indicating that participants in
the clinical group reported more benefits to their eating behaviors, more drawbacks, and a
greater awareness of the function their eating behaviors serve to avoid uncomfortable
emotion. The final significant difference was the EAT-26, which was used to create the
groups.
Just as important as these differences are the measures where there were no
significant differences between the groups. Individuals in the clinical and non-clinical
groups had statistically equivalent scores on the SATAQ (F[1, 247] = 1.69, p = .195) and
the ATSPPHS (F[1, 247] = 0.08, p = .779), suggesting that women’s reported
internalization of societal beauty norms and their attitudes toward counseling were not
different for women above and below the clinical disordered eating level.
89
I also examined mean differences between women who responded “yes” and “no”
to each help seeking question. Women who responded “yes” to the question of whether
or not they wanted to sign up for a counseling appointment reported significantly lower
ATSPPHS scores (F[1, 247] = 6.69, p = .01) as well as higher DBS-Burdens (F[1, 247] =
13.16, p < .001) and DBS-Functional Avoidance Scores (F[1, 247] = 8.49, p = .004) than
women who reported “no,” as can be seen in Table 3. Additionally, women who
responded “yes” to the question of whether or not they wanted to be directed to online
information about eating issues reported significantly higher CORE scores (F[1, 247] =
8.54, p = .004), DBS-Benefits scores (F[1, 247] = 5.88, p = .016), DBS-Burdens scores
(F[1, 247] = 29.84, p < .001), and DBS-Functional Avoidance scores (F[1, 247] = 17.78,
p < .001) than women who responded “no” to this item, as can be seen in Table 4.
90
Chapter 5. Discussion
I will first discuss, interpret, and evaluate the results of the present study,
examining the study as a whole and then each hypothesis separately. Next, I will discuss
the results of additional analyses performed including differences between the clinical
and non-clinical groups. After this, I will discuss limitations of this study including
recommendations for future research. I will also explore the implications of the results of
this study on disordered eating and help-seeking and will discuss possible applications of
this information.
Hypotheses
Hypotheses of the present study aimed to examine help seeking for an eating
disorder by looking at three different dependent variable constructs: attitudes toward
counseling, whether or not a participant wanted online information about disordered
eating, and whether or not a participant wanted to set up a counseling appointment to
discuss food and body image concerns. Specifically, hypotheses were designed to
examine moderators of the relationship between disordered eating behavior and
willingness to seek help, after controlling for general psychological distress. Overall,
most hypotheses were not supported. That is, the five constructs that were hypothesized
to moderate the relation between disordered eating behaviors and willingness to seek help
did not function in the expected way. A significant challenge in the present study was the
small percentage of women who endorsed a desire to set up a counseling appointment
(3%) or obtain online information about disordered eating (12%). This imbalance in the
results likely contributed to our inability to reject several of the null hypotheses.
91
Additionally, no construct accounted for even a moderate amount of the variance
in the criterion variable, help seeking, indicating that help-seeking was a particularly
difficult behavior to predict with the constructs chosen. Also, the suggestion that the five
moderator variables would have predictive power was based on previous studies that
have sampled individuals already seeking help (Cachelin et al., 2000; Cachelin, StriegelMoore, & Regan, 2006; Keel et al., 2002; Goodwin & Fitzgibbon, 2002; Marcos et al.,
2007; Stockford, Turner, & Cooper, 2007). The majority of participants in the present
study were not in counseling and reported a lower severity level than women in most of
the previous studies that have examined eating concerns. The fact that the present results
are different from what was hypothesized may indicate that one cannot generalize from
women with diagnosed eating disorders in a clinic to those who only have mild to
moderate eating concerns and are high functioning.
For all five hypotheses predicting participants’ attitudes towards counseling,
hierarchical multiple regression was used. Each hypothesis involved the same first step:
general psychological distress as a control variable. This first step was not significant (see
Table 7 for these results), indicating that general psychological distress did not
significantly predict attitudes towards counseling. For all hypotheses predicting whether
or not participants were interested in (1) viewing online information about disordered
eating and (2) setting up a counseling appointment, binary logistic regression was used.
General psychological distress, the control variable, significantly discriminated between
participants who were and were not interested in viewing online information about
disordered eating (see Table 8 for these results), but did not significantly discriminate
between participants who were and were not interested in setting up a counseling
92
appointment (see Table 8). Also important to note is that, while general psychological
distress did significantly discriminate between women who were and were not interested
in viewing online information about eating concerns, this construct only accounted for
3.1% of the variance.
It was hypothesized that individuals with more disordered eating would be more
likely to seek help, as previous literature demonstrated that individuals whose disorders
are more severe (and therefore are experiencing higher distress) are more likely to seek
help (Bland et al., 1997; Wang et al., 2005a). The fact that this base hypothesis was only
supported with one of the three measures of help seeking makes it very difficult to
determine the impact of the five hypothesized moderators. There are a few factors which
might contribute to the present study’s results which go against this expectation. First, for
the present study, the entire sample was currently in college. This means that none of the
participants were experiencing a degree of severity that impacted their functioning
enough to force them to withdraw from school. Another important consideration is
impact of a low- to moderate-level eating problem on an individual. Unlike depression or
anxiety, an eating disorder is a coping mechanism. Women with an eating disorder use
their behaviors to manage sadness, anger, and stress, which means that it subjectively
decreases their distress until the behaviors become severe enough to overwhelm them
(Reif & Reif, 2007).
Another issue to consider is that the ATSPPHS was only significantly correlated
with age and DBS-Burdens, and both of these correlations were small (see Table 4). As
attitudes towards counseling are expected to predict actual help seeking, this result is
problematic. The lack of large, prolific correlations between the ATSPPHS and other
93
constructs likely indicates that this study did not measure the variables that most strongly
predict help-seeking in women with low to moderate disordered eating. Although these
correlations were few, and small, it seems important to note that there was a significant
mean difference on the ATSPPHS between women who did and did not desire to
schedule a counseling appointment (see Table 4). With this information about the sample
and control variable in mind, I will discuss each hypothesis individually.
Hypothesis 1. For all three dependent variables it was hypothesized that, after
controlling for general psychological distress, degree of insight would moderate the
relation between degree of disordered eating behavior and willingness to seek help. This
hypothesis, as it relates to participants’ attitudes towards counseling, was not supported.
It is important to highlight that neither the main effects nor the interaction of these main
effects significantly predicted attitudes about counseling. Similarly, as it relates to
participants’ interest in setting up a counseling appointment, this hypothesis was not
supported. Counter to hypotheses, no model (main effects nor main effects plus the
interaction term) produced significant results.
While the specific hypothesis of insight moderating the relation between
disordered eating behavior and help seeking was not supported as it relates to
participants’ interest in viewing information about disordered eating, there were models
that significantly predicted help seeking. As discussed above, psychological distress
significantly accounted for 3.1% of the variance of whether or not participants wanted to
view online information about disordered eating (see Table 8 for these results). The
models adding in the impact of disordered eating and insight as well as the interaction of
these constructs in addition to these two variables were also statistically significant (see
94
Table 8). Only general psychological distress (CORE) significantly contributed to the
models (see Table 9). Therefore, the hypothesized variables did not have any significant
predictive power above and beyond the predictive power of general psychological
distress. It also seems important to note that, while these models were statistically
significant, at the most the model predicted about 4% of the variance. Clinically, this is
not significant and leaves us with questions about what predicts the other 96% of the
variance in these women’s desire (or lack thereof) to get more information about
disordered eating.
While it was expected that insight would significantly moderate the relation
between eating behaviors and help seeking, this result was not found in the present study
for any of the three measures of help seeking. This outcome could be the result of several
different factors. Primarily, I believe that the difference between the hypothesized results
and actual results are related to the differences in population between the present study
and past studies, which was discussed above as it relates to all hypotheses.
Additionally, the hypothesis that insight would have an impact on help-seeking
behavior was based on prior studies relating insight and stage of change; research
indicates that individuals in the precontemplation stage were more likely to have low
insight into eating concerns while individuals in the contemplation stage were more likely
to have higher levels of insight (Stockford, Turner, & Cooper, 2007). It followed, then,
that individuals in a more advanced stage of change would be more likely to seek help.
Perhaps the women in the sample are primarily in a very early stage of change –
precontemplation – and there was not enough variance in this particular characteristic to
see the impact of insight on help-seeking.
95
It also seems important to note here that insight had significant, moderate to large
correlations with several constructs in this study (see Table 4). Participants who reported
greater insight that each of their individual symptoms was connected with an underlying
eating concern also reported many more benefits to the disorder (r = .61), more
drawbacks to the disorder (r = .57), a greater internalization of societal beauty norms (r =
.43), a greater awareness of the function the disorder served to avoid uncomfortable
emotion (r = .39), and more disordered eating behavior (r = .35). These significant
correlations indicate that insight is a powerful variable for women with eating concerns.
Hypothesis 2. For all three dependent variables it was hypothesized that, after
controlling for general psychological distress, internalization of sociocultural beauty
norms would moderate the relation between degree of disordered eating behavior and
willingness to seek help. This hypothesis, as it relates to participants’ attitudes towards
counseling, was not supported. Counter to what was hypothesized, neither the main
effects nor the interaction of these main effects significantly predicted attitudes about
counseling. Similarly, as it relates to participants’ interest in setting up a counseling
appointment, this hypothesis was not supported. No model (main effects nor main effects
plus the interaction term) produced significant results.
While the specific hypothesis of internalization of sociocultural beauty norms
moderating the relation between disordered eating behavior and help seeking was not
supported as it relates to participants’ interest in viewing information about disordered
eating, two of the three models did significantly predict help seeking. As discussed
above, psychological distress significantly predicted whether or not participants wanted
to view online information about disordered eating (see Table 12 for these results). The
96
model adding in the impact of disordered eating and internalization of sociocultural
beauty norms was also statistically significant (see Table 12). The third model, which
added in the interaction of these two main effects, however, was not statistically
significant (see Table 12). Analyses performed to determine what variables significantly
contributed to the third model indicate that only general psychological distress had a
significant contribution (see Table 13). Therefore, it is determined that the hypothesized
variables did not have any significant predictive power above and beyond the predictive
power of general psychological distress.
While it was expected that internalization of sociocultural beauty norms would
significantly moderate the relation between eating behaviors and help seeking, this result
was not found in the present study for any of the three measures of help seeking. This
outcome could be the result of several different factors. Primarily, I believe that the
difference between the hypothesized results and actual results are related to the
differences in population between the present study and past studies, which was
discussed above as it relates to all hypotheses.
Additionally, the hypothesis that internalization of sociocultural beauty norms
would have an impact on help-seeking behavior was based on prior studies relating
internalization of sociocultural beauty norms and problematic eating behaviors; there is
clear evidence in the literature that high internalization of cultural beauty norms is
positively related with body dissatisfaction and negative eating behaviors (Forbes, Jobe,
& Revak, 2006; Morrison & Sheahan, 2009; Rukavina & Pokrajac-Bulian, 2006). In the
present study there is a small, significant correlation between internalized socioucultural
beauty norms and disordered eating (r = .17, p < .01). It was expected that individuals
97
who have personally internalized society’s message that one must be thin, attractive, and
athletic will perceive their eating attitudes and behaviors as more consistent with
society’s expectations and thus will have less interest in seeking help to change these
behaviors than individuals who believe that their behaviors are inconsistent with society’s
messages. Attitudes towards seeking help and internalization of sociocultural beauty
norms were not significantly correlated in the present study.
It also seems important to note here that insight had significant, moderate to large
correlations with several constructs in this study (see Table 4). Participants who reported
greater internalization of sociocultural beauty norms also reported: greater insight that
their individual symptoms are related to an underlying disorder (r = .43), more benefits to
the disorder (r = .41), and more awareness of burdens of the disorder (r = .29).
Hypothesis 3. It was hypothesized that, after controlling for general
psychological distress, perceived benefits of the disorder would moderate the relation
between degree of disordered eating behavior and willingness to seek help for each of
three dependent variables. This hypothesis, as it relates to participants’ attitudes towards
counseling, was not supported. Counter to what was expected, neither the main effects
nor the interaction of these main effects significantly predicted attitudes about counseling
(see Table 15 for results). Similarly, as it relates to participants’ interest in setting up a
counseling appointment, this hypothesis was not supported. No model (main effects nor
main effects plus the interaction term) produced significant results (see Table 16 for
results).
While the specific hypothesis that perceived benefits of the disorder would
moderate the relation between disordered eating behavior and help seeking was not
98
supported as it relates to participants’ interest in viewing information about disordered
eating, all three models did significantly predict help seeking (see Table 16 for these
results). As discussed above, general psychological distress significantly predicted
whether or not participants wanted to view online information about disordered eating,
and the model adding in the impact of disordered eating and perceived benefits of the
disorder was also statistically significant (see Table 16). Finally, the third model, which
added in the interaction of these two main effects, was also statistically significant (see
Table 16). Analyses performed to determine the variables that contributed to the third
significant model indicated that only general psychological distress had a significant
contribution (see Table 17). Therefore, it is determined that the hypothesized variables
did not have any significant predictive power above and beyond the predictive power of
general psychological distress. While these models were statistically significant, their
clinical significance is negligible, as the full model only predicted 5.4% of the variance.
While it was expected that benefits of the disorder would significantly moderate
the relation between eating behaviors and help seeking, this result was not found in the
present study for any of the three measures of help seeking. This outcome could be the
result of several different factors. Primarily, I believe that the difference between the
hypothesized results and actual results are related to the differences in population
between the present study and past studies, which was discussed above as it relates to all
hypotheses.
Additionally, the hypothesis that benefits of the disorder would have an impact on
help-seeking behavior was based on research demonstrating that movement from the
contemplation to the action stage of change is marked by a decrease in reported benefits
99
of the disorder (Prochaska, 1994). It followed, then, that individuals in a more advanced
stage of change will be more likely to seek help. Perhaps the women in the sample are
primarily in a very early stage of change – precontemplation or contemplation – and there
was not enough variance in this particular characteristic to see the impact of perceived
benefits of the disorder on help-seeking.
It also seems important to note here that perception of benefits of the disorder had
significant, moderate to large correlations with several constructs in this study (see Table
4). Participants who reported greater benefits to the disorder also reported greater insight
that their individual symptoms are related to an underlying disorder (r = .61), more
drawbacks to the disorder (r = .58), a greater awareness of the function the disorder
serves to avoid uncomfortable emotion (r = .44), greater internalized societal beauty
norms(r = .41), more general psychological distress (r = .33), and more disordered eating
behavior (r = .32). These correlations seem to provide strong evidence that perceived
benefits of the disorder are an important variable to examine when trying to understand
disordered eating.
Hypothesis 4. It was hypothesized that, after controlling for general
psychological distress, perceived drawbacks of the disorder would moderate the relation
between degree of disordered eating behavior and willingness to seek help for each of
three dependent variables. Perceived drawbacks of the disorder did contribute to a model
which significantly predicted attitudes towards counseling, however the moderation
hypothesis was not supported. While general psychological distress, the control variable,
did not produce a statistically significant model, adding in both disordered eating
behavior and perceived drawbacks of the disorder did significantly predict participants’
100
attitudes towards counseling (see Table 19). This model, which included general
psychological distress, disordered eating behavior, and perceived burdens predicted 3%
of the variance in attitudes toward counseling. The interaction of these two main effects
was added into the third model and this again was statistically significant; it predicted 4%
of the variance in attitudes toward counseling. Adding in the interaction, however, did not
produce a significant change in the percent of variance predicted (see Table 19). While
these models are statistically significant, predicting 4% of the variance is not clinically
significant; clearly many other factors impact a woman’s perceptions of counseling.
As it relates to participants’ interest in setting up a counseling appointment, the
hypothesis that perceived drawbacks of the disorder would moderate the relation between
disordered eating behavior and help seeking was not supported. No model (main effects
nor main effects plus the interaction term) produced significant results (see Table 21 for
results).
While the specific hypothesis that perceived benefits of the disorder would
moderate the relation between disordered eating behavior and help seeking was not
supported as it relates to participants’ interest in viewing information about disordered
eating, all three models did significantly predict help seeking (see Table 21 for these
results). As discussed above, general psychological distress significantly predicted
whether or not participants wanted to view online information about disordered eating,
and the model adding in the impact of disordered eating and perceived benefits of the
disorder was also statistically significant (see Table 21). Finally, the third model, which
added in the interaction of these two main effects, was also statistically significant (see
Table 21). Analyses performed to determine what variables contributed to this third
101
significant model indicated that the main effect of perceived drawbacks of the disorder
was the only significant contributor to the model (see Table 22). Therefore, it is
determined that the interaction did not have any significant predictive power above and
beyond the predictive power of the main effect of perceived drawbacks of the disorder.
Of all the variables tested as moderators of the relation between disordered eating
behavior and help seeking, perceived burdens of the disorder predicted the largest amount
of variance. The full model predicted about 10.9% of the variance in whether or not an
individual wanted to get more information about disordered eating.
While it was expected that burdens of the disorder would significantly moderate
the relation between eating behaviors and help seeking, this result was not found in the
present study for any of the three measures of help seeking. This outcome could be the
result of several different factors. Primarily, I believe that the difference between the
hypothesized results and actual results are related to the differences in population
between the present study and past studies, which was discussed above as it relates to all
hypotheses.
Additionally, the hypothesis that burdens of the disorder would have an impact on
help-seeking behavior was based on research demonstrating that movement from the
precontemplation to the contemplation stage of change is marked by a noted increase in
burdens of the disorder (Prochaska, 1994). Moreover, one study reported that individuals
in the contemplation and action stages reported greater burdens of their eating disorder
than individuals in the precontemplation stage (Cockell, Gellar, & Linden, 2003). It
followed, then, that individuals in a more advanced stage of change will be more likely to
seek help. Perhaps the women in the sample are primarily in a very early stage of change
102
– precontemplation – and there was not enough variance in this particular characteristic
to see the impact of perceived benefits of the disorder on help-seeking.
It also seems important to note here that perception of burdens of the disorder had
significant, moderate to large correlations with several constructs in this study (see Table
4). These correlations seem to provide strong evidence that perceived burdens of the
disorder are an important variable to examine when trying to understand disordered
eating.
Hypothesis 5. It was hypothesized that, after controlling for general
psychological distress, awareness of the function the disorder serves to avoid
uncomfortable emotion (functional avoidance) would moderate the relation between
degree of disordered eating behavior and willingness to seek help for each of three
dependent variables. This hypothesis was not supported for any of the measures of helpseeking. It is important to highlight that neither the main effects nor the interaction of
these main effects significantly predicted attitudes about counseling (see Table 24 for
results). Similarly, as it relates to participants’ interest in setting up a counseling
appointment, this hypothesis was not supported. No model (main effects nor main effects
plus the interaction term) produced significant results (see Table 25 for results).
While the specific hypothesis that functional avoidance would moderate the
relation between disordered eating behavior and help seeking was not supported as it
relates to participants’ interest in viewing information about disordered eating, all three
models did significantly predict help seeking (see Table 25 for these results). As
discussed above, general psychological distress significantly predicted whether or not
participants wanted to view online information about disordered eating, and the model
103
adding in the impact of disordered eating and functional avoidance was also statistically
significant (see Table 25). Finally, the interaction, which added in the interaction of these
two main effects, was also statistically significant (see Table 25). Analyses performed to
determine which variables significantly contributed to this third model indicated that only
the main effect of functional avoidance significantly contributed. Therefore, it is
determined that the interaction did not have any significant predictive power above and
beyond the predictive power of the main effect of functional avoidance. It seems
important to note that the full model (including control variable, degree of disordered
eating, functional avoidance, and the interaction term) predicted 7.2% of the variance in
whether or not a participant wanted to see online information about disordered eating.
While it was expected that functional avoidance would significantly moderate the
relation between eating behaviors and help seeking, this result was not found in the
present study for any of the three measures of help seeking. This outcome could be the
result of several different factors. Primarily, I believe that the difference between the
hypothesized results and actual results are related to the differences in population
between the present study and past studies, which was discussed above as it relates to all
hypotheses.
Additionally, the hypothesis that functional avoidance would have an impact on
help-seeking behavior was based on research demonstrating that individuals in the
contemplation or action stage of change report greater awareness of the function that
eating concerns serve to avoid discomfort than individuals in the precontemplation stage
(Cockell, Gellar, & Linden, 2003). It followed, then, that individuals in a more advanced
stage of change would be more likely to seek help. Perhaps the women in the sample are
104
primarily in a very early stage of change – precontemplation – and there was not enough
variance in this particular characteristic to see the impact of perceived benefits of the
disorder on help-seeking.
It also seems important to note here that functional avoidance had significant,
moderate to large correlations with several constructs in this study (see Table 4). These
correlations seem to provide strong evidence that a woman’s awareness of the function
that the disorder serves to avoid negative emotion is an important variable to examine
when trying to understand disordered eating.
Additional Analyses. Results of the additional analyses indicate that, while
hypotheses of the present study were not supported, there are many important differences
to note within the sample. There were several significant differences between the clinical
and non-clinical groups. Specifically, women who likely met diagnostic criteria for an
eating disorder reported more general psychological distress, greater insight that each of
their symptoms is related to one underlying disorder, greater awareness of the function
their eating behaviors serve to help them avoid uncomfortable emotion, and more
benefits and drawbacks to their behaviors than women who did not meet clinical criteria.
This seems to indicate that the women in this sample with more disordered eating
behavior (who, notably, are still functioning college students) have quite a bit of insight
into their behaviors in general.
Equally important are the two constructs where there was no significant
difference between women above and below the clinical cutoff: internalization of societal
beauty norms and attitudes towards counseling. It seems that women who have moderate
to high disordered eating are no more likely than women with only mild disordered eating
105
behavior to have internalized strong societal messages about beauty. Additionally, these
two groups of women have statistically equivalent attitudes toward counseling. This
result may contribute to the lack of significance in the hypotheses. If women who have
more disordered eating behavior have equivalent views of counseling as women with less
disordered eating behavior, it follows that they will be no more (and no less) likely to
seek therapy.
In addition to examining mean differences of women above and below the clinical
level, I also examined mean differences between women who said “yes” and “no” to each
help seeking question. Women who indicated a desire to set up a counseling appointment
to discuss food and body image issues reported significantly lower stigma (more positive
attitudes) towards counseling, more drawbacks to their disorder, and a greater awareness
of the function their eating behaviors serve to avoid uncomfortable emotion than women
who indicated no interest in a counseling appointment. These results are consistent with
prior research, in that there was a relation between attitudes towards counseling and
reported interest in setting up an appointment (Vogel, et al., 2007). Additionally, prior
research demonstrated that there is a connection between an increase in reported
drawbacks to the disorder and the movement from the precontemplation to the
contemplation stage of change (Cockell, Gellar, & Linden, 2003). It was originally
hypothesized that this movement from one stage to the next would coincide with a greater
willingness to seek help. Finally, women who have more awareness that their eating
behaviors are connected to their emotions – and not just the result of beliefs about their
body or food – are likely to be more open to counseling, where emotions are often the
focus.
106
Finally, women who responded “yes” to the question of whether or not they
wanted to be directed to online information about eating issues, when compared with
women who responded “no” to this question, reported more general psychological
distress, more benefits and drawbacks to their behavior, and a greater awareness of the
function their behavior serves to help them avoid uncomfortable emotion. These results
regarding benefits of the disorder and awareness that the disorder helps them avoid
uncomfortable emotions are similar to the results discussed above, for women who
expressed interest in a counseling appointment, and likely for similar reasons.
Additionally, the result that women who were more interested in learning about eating
issues also reported experiencing more general psychological distress is consistent with
prior research which indicates that individuals with more distress are more likely to seek
counseling (Bland et al., 1997; Wang et al., 2005a).
Limitations and Direction for Future Research
As hypotheses in the present study were not supported, it seems quite important to
examine limitations. One of the primary limitations in the present study is the percentage
of women who reported an interest in seeking counseling or getting online information
about disordered eating. This could have had a lesser impact if the overall sample was
larger. As the present study was performed because women with eating concerns so
infrequently seek help, this finding is not surprising. In future studies it may prove
beneficial to assess different levels of help seeking (present or past counseling, interest in
seeking counseling in the future, having looked for information or support online,
speaking with a parent or friend about their eating concerns, etc) and different factors that
may impact a woman’s perception of need (“has a doctor, family member, or friend
107
expressed concern about eating/exercise?”, “have you visited any websites about eating
and exercise concerns?”). As it appears that the base rate is so low, another possible
solution could be to increase the sample size significantly such that the group of women
who report interest in seeking counseling is 5-6 times larger than the present group.
Additionally, it was assumed that women in the current sample would have been
dispersed normatively across the Stages of Change (Prochaska & DiClimente, 1985). For
this reason, this variable was not measured. It seems likely, given the very few women
who expressed interest in help seeking, that most of these women were in the precontemplative stage and therefore unaware of having a problem.
Another important factor of this study, the population, could be perceived as a
limitation. Literature that was used to determine the present hypotheses typically
involved women who had diagnosable eating disorders and were (willingly or
unwillingly) seeking treatment. Because the present study tested hypotheses suggested by
research on a more severe population with a group of functioning college women, we
cannot be certain why the hypotheses are rejected. Perhaps the moderator effects that
were suggested simply don’t exist in any population; conversely they may exist in a more
severe population, but not in the present, functional college population. Because of this
uncertainty, the present population could be a limitation (perhaps their level of severity
and functioning made it impossible to see an existing effect that would be found with
women with eating disorders), or perhaps this population is not a limitation, but just
provides the field with information about women with lower severity eating concerns.
For this reason, future research could test similar hypotheses with a group of
women who have more severe struggles with disordered eating. This could require a very
108
large sample, where women without diagnosable eating behaviors are screened out.
These hypotheses could also be tweaked to be tested with women who are attending a
first counseling appointment. In this case, women’s perception of their need for
counseling – and the degree to which they are coming willingly rather than through
coercion – would be the dependent variables. Also important to note is that many prior
studies were conducted in person, while the present study was conducted online. The
format of the assessment could have both benefits (anonymity related to truthfulness,
taken in a comfortable environment) and drawbacks (no connection with a clinician,
amount of energy/focus given to measures).
It seems important to note that general psychological distress did significantly
predict membership of women who did and did not want to get more information about
disordered eating. It seems equally important to note that level of psychological distress
predicted less than 4% of the variance. Future research could work toward understanding
the help-seeking constructs examined in the present study to gain a better sense for what
best predicts help-seeking behavior for women with eating concerns.
As few constructs in the present study were correlated with participants attitudes
towards counseling (ATSPPHS), it seems important to consider other constructs that may
predict help seeking for women with mild to moderate disordered eating. Perhaps factors
outside of the individual’s perception of need predicts help seeking – constructs such as
Prochaska and DiClemente’s Stages of Change, the individual’s expectation/hope that
therapy can help (and not just take away their coping mechanism), and perceived cause of
the disorder may prove useful in future studies.
109
Finally, some of the consequences of this line of research being relatively new
may have been limitations. Hypotheses in the present study were drawn from a relatively
small body of literature, with very few studies on actual help seeking behavior. In this
study I attempted to put together several lines of research – stages of change, perceptions
of need, and specific help seeking – without having literature that provided evidence of
each relationship. An additional factor related to the youth of this line of study is the
measures that were used. While all measures used demonstrated good internal validity (in
past studies and the present study), most have not been tested with a wide range of
populations, especially the measures used to quantify hypothesized moderator constructs.
Implications
Implications of the present study are varied and can be drawn from the rejected
hypotheses as well as mean differences between groups. Descriptive statistics again
demonstrate that women with eating concerns are not likely to seek counseling. However,
women who indicated that they wanted to set up a counseling appointment reported
significantly lower stigma towards counseling. This again supports the importance of
stigma in seeking therapy, suggesting that minimizing stigma towards counseling may
help women with disordered eating to seek therapy. Additionally, these women reported
significantly more drawbacks to the disorder and a greater awareness of the function their
disorder serves to avoid uncomfortable emotion. This suggests that outreach could be
developed to help women explore and better understand their disorder – especially the
aspects of their eating behaviors that get in the way of being happy, productive, and
successful – with the expectation that these activities may move them towards greater
willingness to seek help.
110
In examining the five hypotheses, the most notable result is that these constructs
did not predict help seeking as they were expected to and predicted very little of the
variance. What this suggests is that there is still much to understand about help seeking
for an eating disorder and that women’s choices to seek help for disordered eating are
primarily related to something other than general psychological distress, disordered
eating behaviors, and the five hypothesized moderators.
The results of the present study also give us a better picture of characteristics of
high functioning college women with eating concerns. Women who are at a clinical level
of disordered eating report more general psychological distress than women below the
clinical level as well as greater reported insight that each separate symptom they
experience is related to one underlying disorder. Additionally, these women report more
benefits to the disorder, more drawbacks, and a greater awareness that their eating
behaviors serve a function of avoiding uncomfortable emotion. This level of awareness
indicates that the women who likely meet clinical criteria for an eating disorder have
more insight into their behavior than women who do not meet clinical criteria. These
results also seem to demonstrate that, as disordered eating behaviors increase, both the
benefits and the drawbacks of these behaviors increase and that women who likely meet
clinical criteria are aware of these increases.
Finally, results of the present study validate some existing (but relatively new)
measures. The measure used in the present study to assess degree of insight that disparate
symptoms are related to one underlying disorder was developed in 2007 (Stockford,
Turner, & Cooper) and has not been commonly cited in recent literature. The present
study suggests that this measure has good internal consistency and that the construct is an
111
important one to examine for disordered eating, as this instrument was significantly
correlated with most other constructs in this study (see Table 4). The same can be said for
the Decisional Balance Scale, which was modified to assess eating disorders in 2002 by
Cockell, Gellar, and Linden. In the present study, the subscales of this measure have good
internal consistency and are significantly correlated with most other instruments,
including those that are well-established in the disordered eating literature.
112
References
Al Anbar, N. N., Dardennes, R. M., Prado-Netto, A., Kaye, K., Contejean, Y. (2010).
Treatment choices in autism spectrum disorder: The role of parental illness
perceptions. Research in Developmental Disabilities, 31, 817-828.
Alvarez-Rayon, G., Mancilla-Diaz, J. M., Vazquez-Arevalo, R., Unikel-Santoncini, C.,
Caballero-Romo, A., & Mercado-Corona. (2004). Validity of the Eating Attitudes
Test: A study of Mexican eating disorder patients. Eating and Weight Disorders,
9, 243-248.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental
disorders (Revised 4th ed.). Washington, DC: Author.
Anderson, R. M. (1995). Revisiting the behavioral model and access to medical care:
Does it matter? Journal of Health and Social Behavior, 36, 1-10.
Anderson, R. M. & Newman, J. F. (1973). Societal and individual determinants of
medical care utilization in the United States. Milbank Memorial Fund Quarterly
Journal, 51, 95-124.
Andrews, G., Singh, M., & Bond, M. (1993). The Defense Style Questionnaire. Journal
of Nervous and Mental Disease, 181, 246-254.
Banasiak, S. J., Wertheim, E. H., Koerner, J., & Voudouris, N. J. (1999). Test-retest
reliability and internal consistency of a variety of measures of dietary restraint and
body concerns in a sample of adolescent girls. International Journal of Eating
Disorders, 29, 85-89.
Bedford, J. L. & Johnson, C. S. (2006). Societal influences on body image dissatisfaction
in younger and older women. Journal of Women and Aging, 18, 41-55.
113
Bell, M. D., Billington, R., & Becker, B. (1986). A scale for the assessment of object
relations: Reliability, validity, and factorial invariance. Journal of Clinical
Psychology, 42, 733-741.
Bergeron, E., Poirier, L. R., Fournier, L. Roberge, P., & Barrette, G. (2005).
Determinants of service use among young Canadians with mental disorders.
Canadian Journal of Psychiatry, 50, 629-636.
Betz, N. (1987). Use of discriminant analysis in counseling psychology research. Journal
of Counseling Psychology. Special Issue: Quantitative foundations of counseling
psychology research, 34, 393-403.
Blake, W., Turnbull, S., & Treasure, J. L. (1997). Stages and processes of change in
eating disorders; Implications for therapy. Clinical Psychology and
Psychotherapy, 4, 186-191.
Bland, R. C., Newman, S. C., & Orn, H. (1997). Help-seeking for psychiatric disorders.
Canadian Journal of Psychiatry, 42, 935-942.
Byrne, S. M., Allen, K., L., Lampard, A. M., Dove, E. R., & Furstland, A. (2009). The
factor structure of the Eating Disorder Examination in clinical and community
samples. International Journal of Eating Disorders, 43, 260-265.
Cachelin, F. M., Rebeck, R., Veisel, C., & Striegel-Moore, R. H. (2000). Barriers to
treatment for eating disorders among ethnically diverse women. International
Journal of Eating Disorders, 30, 269-278.
Cachelin, F. M., Striegel-Moore, R. H., & Regan, P. C. (2006). Factors associated with
treatment seeking in a community sample of European American and Mexican
114
American women with eating disorders. European Eating Disorders Review, 14,
422-429.
Carragher, N., Adamson, G., Bunting, B., & McCann, S. (2010). Treatment-seeking
behaviours for depression in the general population: Results from the National
Epidemiologic Survey on Alcohol and Related Conditions. Journal of Affective
Disorders, 121, 59-67.
Cockell, S. J., Geller, J., & Linden, W. (2003). Decisional balance in Anorexia Nervosa:
Capitalizing on ambivalence. European Eating Disorders Review, 11, 75-89.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Deane, F. P. & Chamberlain, K. (1994). Treatment fearfulness and distress as predictors
of professional psychological help-seeking. British Journal of Guidance and
Counseling, 22, 207-217.
Deane, F. P. & Todd, D. M. (1996). Attitudes and intentions to seek professional
psychological help for personal problems or suicidal thinking. Journal of College
Student Psychotherapy, 10, 45-59.
Fava, J. L., Velicer, W. F., & Prochaska, J. O. (1995). Applying the transtheoretical
model to a representative sample of smokers. Addictive Behaviors, 20, 189-203.
Fairburn, C. G., & Cooper, Z. (1993). The eating disorder examination (12th ed.). In C. G.
Fairburn, & G. T. Wilson (Eds.), Binge eating: Nature, assessment, and treatment
(pp. 317-360). New York: Guilford Press.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (2001). Structured Clinical
Interview for DSM-IV-TR Axis I Disorders – Nonpatient edition. New York:
Biometrics Research Department, New York State Psychiatric Institute.
115
Fischer, E. H. & Farina, A. (1995). Attitudes toward seeking professional psychological
help: A shortened form and considerations for research. Journal of College
Student Development,36, 368-373.
Fischer, E. H., & Turner, J. I. (1970). Orientations to seeking professional help:
Development and research utility of an attitude scale. Journal of Consulting and
Clinical Psychology, 35, 79-90.
Forbes, G. B., Jobe, R. L., & Revak, J. A. (2006). Relationships between dissatisfaction
with specific body characteristics and the Sociocultural Attitudes Toward
Appearance Questionnaire-3 and Objectified Body Consciousness Scale. Body
Image, 3, 295-300.
Fredrickson, B., & Roberts, T. (1997). Objectification theory: Toward understanding
women’s lived experiences and mental health risks. Psychology of Women
Quarterly, 21, 173-178.
Garner, D. M. (2004). Eating Disorder Inventory-3: Professional Manual. Lutz, FL:
Psychological Assessment Resources, Inc.
Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: An index of the
symptoms of anorexia Nervosa. Psychological Medicine, 9, 273-279.
Garner, D. M., Olmsted, M. (1984). Eating Disorder Inventory Manual. New York:
Psychological Assessment Resources.
Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982). The Eating Attitudes
Test: Psychometric features and clinical correlates. Psychological Medicine, 12,
871-878.
116
Goodwin, R. D., Fitzgibbon, M. L. (2002). Social anxiety as a barrier to treatment for
eating disorders. International Journal of Eating Disorders, 32, 103-106.
Griffiths, R. A., Beumont, P. J. V, Russell, J., Schotte, D., Thornton, C., Touyz, S.,
Varano, P. (1999). Sociocultural attitudes towards appearance in dieting
disordered and nondieting disordered subjects. European Eating Disorders
Review, 7, 193-203.
Griffiths, R. A., Mallia-Blanco, R., Boesenberg, E., Ellis, C., Fischer, K., Taylor, M., &
Wyndham, J. (2000). Restrained eating and sociocultural attitudes to appearance
and general dissatisfaction. European Eating Disorders Review, 8, 394-402.
Grimley, D. M., Riley, G. E., Bellis, J. M., & Prochaska, J. O. (1993). Assessing the
stages of change and decision making for contraceptive use for the prevention of
pregnancy, sexually transmitted diseases, and acquired immunodeficiency
syndrome. Health Education Quarterly, 20, 455-470.
Harman, J. S., Edlund, M. J., & Fortney, J. C. (2004) Disparities in the adequacy of
depression treatment in the United States. Psychiatric Services, 55, 1379-1385.
Heinberg, L. J., Thompson, J. K., & Stormer, S. (1995). Development and validation of
the Sociocultural Attitudes Towards Appearance Questionnaire (SATAQ).
International Journal of Eating Disorders, 17, 81-89.
Hepworth, N. & Paxton, S. J. (2007). Pathways to help-seeking in Bulimia Nervosa and
Binge Eating problems: A concept mapping approach. International Journal of
Eating Disorders, 40, 493-504.
117
Herman, C. P., Polivy, J., Pliner, P., Threkheld, J., Munic, J. (1978). Distractability in
dieters and non dieters: An alternative view of ‘Externality.’ Journal of
Personality and Social Psychology, 36, 536-548.
Hohlstein, L. A., Smith, G. T., & Atlas, J. G. (1998). An application of expectancy theory
to eating disorders: Development and validation of measures of eating and dieting
expectancies. Psychological Assessment, 10, 49-58.
Holliday, J., Wall, E., Treasure, J., Weinman, J. (2005). Perceptions of illness in
individuals with Anorexia Nervosa: A comparison with lay men and women.
International Journal of Eating Disorders, 37, 50-56.
Jackson, H., Judd, F., Komiti, A., Fraser, C., Murray, G., Robins, G., Pattison, P., &
Wearing, A. (2007). Mental health problems in rural contexts: What are the
barriers to seeking help from professional providers? Australian Psychologist, 42,
147-160.
Karasz, A. (2005). Cultural differences in conceptual models of depression. Social
Science & Medicine, 60, 1625-1635.
Kandel, D. B. (1980). Drug and drinking behavior among youth. Annual Review of
Sociology, 6, 235-285.
Keel, P. K., Dorer, D. J., Eddy, K. T., Delinsky, S. S., Franko, D. L., Blais, M. A., …
Herzog, D. B. (2002). Predictors of treatment utilization among women with
anorexia and bulimia Nervosa. The American Journal of Psychiatry, 159, 140142.
Kjelsas, E., Bjornstrom, C., & Gotestam, K. G. (2003). Prevalence of eating disorders in
female and male adolescents (14-15 years). Eating Behaviors, 5, 13-25.
118
Kushner, M. G., & Sher, K. J. (1989). Fear of psychological services and its relation to
service utilization. Professional Psychology: Research and Practice, 20, 251-257.
Lawson, V. L., Bundy, C., Lyne, P. A., & Harvey, J. N. (2004). Using the IPQ and PMDI
to predict regular diabetes care-seeking among patients with Type 1 diabetes.
British Journal of Health Psychology, 9, 241-252.
Leventhal, H. Nerenz, D. R., & Steele, D. J. (1984). Illness representations and coping
with health threats. In A. Baum, and S. E. Taylor and Je. E. Singer (Eds.).
Handbook of Psychology and Health, Volume IV: Social Psychological Aspects of
Health. Pp. 219-252. New York: Wiley.
Marcos, Y. Q., Cantero, M. C. T., Escobar, C. R., Acosta, G. P. (2007). Illness perception
in eating disorders and psychosocial adaptation. European Eating Disorders
Review, 15, 373-384.
McKinley, N. M. & Hyde, J. S. (1996). The Objectified Body Consciousness Scale
development and validation. Psychology of Women Quarterly, 20, 181-215.
Mendelson, B. K., Mendelson, M. J., & Andrews, J. (2001). Body-esteem scale for
adolescents and adults. Journal of Personality Assessment, 76, 90-106.
Mintz, L. B. & O’Halloran, M. S. (2000). The Eating Attitudes Test: Validation with
DSM-IV eating disorder criteria. Journal of Personality Assessment, 74, 489-503.
Miranda, J. & Cooper, L. A. (2004). Disparities in care for depression among primary
care patients. Journal of General Internal Medicine, 19, 120-126.
Mitchell, K. S. & Mazzeo, S. E. (2009). Evaluation of a structural model of
objectification theory and eating disorder symptomatology among European
119
American and African American undergraduate women. Psychology of Women
Quarterly, 33, 384-395.
Morrison, T. G. & Sheahan, E. E. (2009). Gender-related discourses as mediators in the
association between internalization of the thin-body ideal and indicants of body
dissatisfaction and disordered eating. Psychology of Women Quarterly, 33, 374383.
Morry, M. M. & Staska, S. L. (2001). Magazine exposure: Internalization, selfobjectification, eating attitudes, and body satisfaction in male and female
university students. Canadian Journal of Behavioural Science, 33, 269-279.
O’Connell, D. & Velicer, W. F. (1988). A decisional balance measure and the stages of
change model for weight loss. International Journal of the Addictions, 23, 729750.
Parslow, R. A., & Jorm, A. F. (2000). Who uses mental health services in Australia? An
analysis of data from the National Survey on Mental Health and Well-Being.
Australian and New Zealand Journal of Psychiatry, 34, 997-1008.
Petrie, T., Rogers, R., Johnson, C., & Diehl, N. (1996, August). Development and
validation of the Beliefs About Attractiveness Scale-Revised. Poster presented at
the 104th annual convention of the American Psychological Association, Toronto,
Ontario, Canada.
Prochaska, J. O. (1994). Strong and weak principles for progressing from
precontemplation to action on the basis of twelve problem behaviors. Health
Psychology, 13, 47-51.
120
Prochaska, J. O., & DiClemente, C. C. (1985). Common processes of change in smoking,
weight control, and psychological distress. In S. Shiffman, & T. Wills (Eds.),
Coping and substance abuse (pp. 345-363). San Diego, CA: Academic Press.
Prochaska, J. O., Velicer, W. F., Rossi, J. S., Goldstein, M. G., Marcus, B. H., Rakowski,
W., … Rossi, S. R. (1994). Stages of change and decisional balance for 12
problem behaviors. Health Psychology, 11, 111-118.
Prouty, A. M., Protinsky, H. O., & Canady, D. (2002). College women: Eating behaviors
and help-seeking preferences. Adolecence, 37, 353-363.
Reif, D. and Reiff, K. (1997). Eating Disorders: Nutrition Therapy in the Recovery
Process. Mercer Island, WA: Life Enterprises.
Rhodes, A. E., Goering, P. N., To, T., & Williams, J. I. (2002). Gender and outpatient
mental health service use. Social Science & Medicine, 54, 1-10.
Rimer, B. K., Conaway, M. R., Lyna, P. R., & Rakowski, W. (1996). Cancer screening
practices among women in a community health center population. American
Journal of Preventative Medicine, 12, 351-357.
Rukavina, T. & Pokrajac-Bulian, A. (2006). Thin-ideal internalization, body
dissatisfaction and symptoms of eating disorders in Croatian adolescent girls.
Eating and Weight Disorders, 11, 31-37.
Sareen, J. Cox, B. J., Afifi, T. O., Yu, B. N., Stein, M. B. (2005). Mental health service
use in a nationally representative Canadian survey. Canadian Journal of
Psychiatry, 50, 753-761.
Shisslak, C. M., Crago, M., & Estes, L. S. (1995). The spectrum of eating disturbances.
International Journal of Eating Disorders, 18, 209-219.
121
Sinclair, A., Barkham, M., Evans, C., Connell, J., & Audin, K. (2005). Rationale and
development of a general population well-being measure: Psychometric status of
the GP-CORE in a student sample. British Journal of Guidance and Counselling,
33, 153-173.
Steinhausen, H. C., Rauss-Mason, C., & Seidel. (1991). Follow-up studies of anorexia
Nervosa: A review of four decades of outcome research. Psychological Medicine,
21, 447-454.
Stockford, K., Turner, H., Cooper, M. (2007). Illness perception and its relationship to
readiness to change in the eating disorders: A preliminary investigation. British
Journal of Clinical Psychology, 46, 139-154.
Stice, E. (2001). A prospective test of the dual pathway model of bulimic pathology:
Mediating effects of dieting and negative affect. Journal of Abnormal Psychology,
110, 124-125.
Stice, E., & Agras, W. S. (1998). Predicting the onset and remission of bulimic behaviors
in adolescence: A longitudinal grouping analysis. Behavior Therapy, 29, 257-276.
Stice, E., Marti, C. N., Shaw, H., & Jaconis, M. (2009). An 8-year longitudinal study of
the natural history of threshold, subthreshold, and partial eating disorders from a
community sample of adolescents. Journal of Abnormal Psychology, 118, 587597.
Stice, E., Schupak-Neuberg, E., Shaw, H. E., & Stein, R. I. (1994). Relation of media
exposure to eating disorder symptomatolgy: An examination of mediating
mechanisms. Journal of Abnormal Psychology, 103, 836-840.
122
Stice, E., Telch, C. F., Rizvi, S. L. (2000). Development and validation of the Eating
Disorder Diagnostic Scale: A brief self-report measure of anorexia, bulimia, and
binge-eating disorder. Psychological Assessment, 12, 123-131.
Sullivan, P. F. (1995). Mortality in anorexia Nervosa. American Journal of Psychiatry,
152, 1073-1074.
ten Have, M., de Graaf, R., Vollebergh, W., & Beekman, A. (2004). Results from the
Netherlands Mental Health Survey and Incidence Study (NEMESIS). Journal of
Affective Disorders, 80, 239-248.
Thompson, J. K., van den Berg, P., Roehrig, M., Guarda, A. S., & Heinberg, L. J. (2004).
The Sociocultural Attitudes Towards Appearance Scale-3 (SATAQ-3):
Development and validation. International Journal of Eating Disorders, 35, 293304.
Tiggemann, M. (2003). Media exposure, body dissatisfaction, and disordered eating:
Television and magazines are not the same! European Eating Disorders Review,
11, 418-430.
Tishby, O., Turel, M., Gumpel, O., Pinus, U., Lavy, S. B., Winokour, M., &
Sznajderman, S. (2001). Help-seeking attitudes among Israeli adolescents.
Adolescence, 36, 249-264.
Vanheusden, K., van der Ende, J., Mulder, C. L., van Lenthe, F. J., Verhulst, F. C., &
Mackenbach, J. P. (2009). Beliefs about mental health problems and help-seeking
behavior in Dutch young adults. Social Psychiatry and Psychiatric Epidemiology,
44, 239-246.
123
Vartanian, L. R. & Hopkinson, M. M. (2010). Social connectedness, conformity, and
internalization of societal standards of attractiveness. Body Image, 7, 86-89.
Vasiliadis, H. M., Lesage, A., Adair, C., Wang, P. S., & Kessler, R. C. (2007) Do Canada
and the United States differ in prevalence of depression and utilization of
services? Psychiatric Services, 58, 63-71.
Velicer, W. F., DiClemente, C. C., Prochaska, J. O., & Brandenburg, N. (1985).
Decisional balance measure for assessing and predicting smoking status. Journal
of Personality and Social Psychology, 48, 1279-1289.
Vogel, D. L., & Wei, M. (2005). Adult attachment and help-seeking intent: The
mediating roles of psychological distress and perceived social support. Journal of
Counseling Psychology, 52, 347-357.
Vogel, D. L. & Wester, S. R. (2003). To seek help or not to seek help: The risks of selfdisclosure. Journal of Counseling Psychology, 50, 351-361.
Vogel, D. L., Wade, N. G., Wester, S. R., Larson, L., & Hackler, A. H. (2007). Seeking
help from a mental health professional: The influence of one’s social network.
Journal of Clinical Psychology, 63, 233-245.
Vogel, D. L., Wester, S. R., Wei, M., Boysen, G. A. (2005). The role of outcome
expectations and attitudes on decisions to seek professional help. Journal of
Counseling Psychology, 52, 459-470.
Wampold, B. E. & Freund, R. D. (1987). Use of multiple regression in counseling
psychology research: A flexible data-analytic strategy. Journal of Counseling
Psychology. Special Issue: Quantitative foundations of counseling psychology
research, 34, 372-382.
124
Wang, J., Patten, S. B., Williams, J. V., Currie, S., Beck, C. A., Maxwell, C. J., & ElGuebaly, N. (2005a). Help-seeking behaviors of individuals with mood disorders.
The Canadian Journal of Psychiatry, 50, 652-659
Wang, P. S., Lane, M., Olfson, M., Pincus, H. A., Wells, K. B., & Kessler, R. C. (2005b).
Twelve-month use of mental health services in the United States: Results from the
NCS-R. Archives of General Psychiatry, 62, 629-640.
Weinman, J., Petrie, K. J., Moss-Morris, R., Horne, R. (1996). The Illness Perception
Questionnaire: A new method for assessing the cognitive representation of illness.
Psychology and Health, 11, 431-445.
Wells, J. E., Robins, L. N., Bushnell, J. A., Jarosz, D. & Oakley-Browne, M. A. (1994).
Perceived barriers to care in St. Louis (USA) and Christchurch (NZ): Reasons for
not seeking professional help for psychological distress. Social Psychiatry and
Psychiatric Epidemiology, 29, 155-164.
Wilson, C. J. & Deane, F. P. (2010). Help-negation and suicidal ideation: The role of
depression, anxiety, and hopelessness. Journal of Youth and Adolescence, 39,
291-305.
Young, A. S., Klap, R., Sherbourne, C. D., & Wells, K. B. (2001). The quality of care for
depressive and anxiety disorders in the US. Archives of General Psychiatry, 58,
55-61.
125
Table 1
Descriptive Statistics for Complete Sample
Variable
Age
SATAQ
ATSPPHS
CORE
IPQ-R
EAT
DBS
Benefits
Burdens
Functional Avoidance
Minimum
Maximum
Mean
18.00
1.27
1.00
1.00
0.00
7.00
43.00
4.93
3.90
4.25
4.00
75.00
19.46
3.61
2.35
2.30
2.68
18.81
1.00
1.00
1.00
5.00
4.67
4.00
.
2.79
2.20
1.60
.
.
Standard
Deviation
2.83
.64
.53
.68
.84
11.12
.
.
.92
.87
.75
Note: N = 249; SATAQ = Sociocultural Attitudes Towards Appearance Questionnaire;
ATSPPHS= Attitudes Towards Seeking Professional Psychological Help Scale; CORE=
Clinical Outcomes in Routine Evaluation – Short Form; IPQ-R= Illness Perception
Questionnaire – Revised; identity subscale; EAT= Eating Attitudes Test 26; DBS=
Decisional Balance Scale. Higher scores on the SATAQ, ATSPPHS, CORE, IPQ-R,
EAT, DBS-Benefits, DBS-Burdens, and DBS-Functional avoidance indicate more
internalization of sociocultural beauty norms, more stigma towards counseling, more
general psychological distress, more insight that individual symptoms are related to an
underlying disorder, more disordered eating behavior, more perceived benefits to the
disorder, more perceived drawbacks to the disorder, and a greater awareness of the
function the disorder serves to avoid uncomfortable emotion, respectively.
126
Table 2
Descriptive Statistics Split by Whether or Not Participant Meets Clinical Criteria
Variable
Age
SATAQ
ATSPPHS
CORE
IPQ-R
EAT
DBS
Benefits
Burdens
Functional
dAvoidance
Clinical
N
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
201
48
201
48
201
48
201
48
201
48
201
48
No
Yes
No
Yes
No
Yes
201
48
201
48
201
48
Minimum
Maximum
Mean
18
18
1.27
2.13
1.00
1.00
1.00
1.08
0.00
.83
7.00
12.81
43.00
23.00
4.93
4.80
3.90
3.90
4.25
4.00
4.00
4.00
25.00
75.00
19.53
19.19
3.58
3.71
2.35
2.33
2.23*
2.62*
2.61*
3.00*
14.60*
36.43*
1.00
1.13
1.00
1.00
1.00
1.00
5.00
5.00
4.13
4.67
4.00
3.57
2.65*
3.39*
2.07*
2.76*
1.49*
2.08*
Standard
Deviation
3.09
1.20
.64
.57
.53
.56
.64
.73
.85
.74
4.96
12.44
.87
.90
.80
.91
.68
.87
Note: An * indicates a significant mean difference at p < .05 level between women who
meet clinical criteria and women who do not; SATAQ = Sociocultural Attitudes Towards
Appearance Questionnaire; ATSPPHS= Attitudes Towards Seeking Professional
Psychological Help Scale; CORE= Clinical Outcomes in Routine Evaluation – Short
Form; IPQ-R= Illness Perception Questionnaire – Revised; identity subscale; EAT=
Eating Attitudes Test 26; DBS= Decisional Balance Scale. Higher scores on the SATAQ,
ATSPPHS, CORE, IPQ-R, EAT, DBS-Benefits, DBS-Burdens, and DBS-Functional
avoidance indicate more internalization of sociocultural beauty norms, more stigma
towards counseling, more general psychological distress, more insight that individual
symptoms are related to an underlying disorder, more disordered eating behavior, more
perceived benefits to the disorder, more perceived drawbacks to the disorder, and a
greater awareness of the function the disorder serves to avoid uncomfortable emotion,
respectively.
127
Table 3
Descriptive Statistics Split by Whether or Not Participant Wanted to View Online Eating
Issues Information
Variable
Age
SATAQ
ATSPPHS
CORE
IPQ-R
EAT
DBS
Benefits
Response
N
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
219
30
219
30
219
30
219
30
219
30
219
30
Minimum
18.00
18.00
1.27
1.63
1.00
1.00
1.00
1.25
0.00
0.00
7.00
8.00
Maximum
Mean
43.00
41.00
4.93
4.80
3.90
2.90
4.25
4.00
4.00
4.00
75.00
55.00
19.39
20.00
3.59
3.72
2.37
2.18
2.26*
2.64*
2.65
2.93
18.64
20.02
Standard
Deviation
2.61
4.10
.62
.68
.53
.50
.66
.74
.82
.92
11.15
10.96
No
219
1.00
5.00
2.74
.94
Yes
30
1.13
4.50
3.17
.68
Burdens
No
219
1.00
4.67
2.10*
.82
Yes
30
1.07
4.47
2.97*
.82
Functional
No
219
1.00
4.00
1.53*
.72
dAvoidance
Yes
30
1.00
3.57
2.13*
.78
Note: An * indicates a significant mean difference at p < .01 level between women who
responded “yes” and “no” when asked whether they wanted to be directed to online
information about eating issues; SATAQ = Sociocultural Attitudes Towards Appearance
Questionnaire; ATSPPHS= Attitudes Towards Seeking Professional Psychological Help
Scale; CORE= Clinical Outcomes in Routine Evaluation – Short Form; IPQ-R= Illness
Perception Questionnaire – Revised; identity subscale; EAT= Eating Attitudes Test 26;
DBS= Decisional Balance Scale. Higher scores on the SATAQ, ATSPPHS, CORE, IPQR, EAT, DBS-Benefits, DBS-Burdens, and DBS-Functional avoidance indicate more
internalization of sociocultural beauty norms, more stigma towards counseling, more
general psychological distress, more insight that individual symptoms are related to an
underlying disorder, more disordered eating behavior, more perceived benefits to the
disorder, more perceived drawbacks to the disorder, and a greater awareness of the
function the disorder serves to avoid uncomfortable emotion, respectively.
128
Table 4
Descriptive Statistics Split by Whether or Not Participant Wanted to be Directed to
Schedule a Counseling Appointment
Variable
Age
SATAQ
ATSPPHS
CORE
IPQ-R
EAT
DBS
Benefits
Response
N
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
242
7
242
7
242
7
242
7
242
7
242
7
Minimum
18.00
18.00
1.27
2.87
1.00
1.00
1.00
2.04
0.00
1.90
7.00
8.00
Maximum
Mean
43.00
41.00
4.93
4.57
3.90
2.30
4.25
3.58
4.00
4.00
75.00
47.00
19.38
22.43
3.61
3.60
2.36*
1.84*
2.29
2.73
2.67
3.07
18.68
23.31
Standard
Deviation
2.51
8.24
.63
.61
.52
.52
.68
.57
.84
.74
11.03
14.08
No
242
1.00
5.00
2.78
.93
Yes
7
1.50
3.50
2.98
.67
Burdens
No
242
1.00
4.67
2.17*
.85
Yes
7
2.40
4.00
3.34*
.61
Functional
No
242
1.00
4.00
1.58*
.75
dAvoidance
Yes
7
1.43
3.29
2.41*
.63
Note: An * indicates a significant mean difference at p < .01 level between women who
responded “yes” and “no” when asked whether they wanted to be directed to schedule a
counseling appointment; SATAQ = Sociocultural Attitudes Towards Appearance
Questionnaire; ATSPPHS= Attitudes Towards Seeking Professional Psychological Help
Scale; CORE= Clinical Outcomes in Routine Evaluation – Short Form; IPQ-R= Illness
Perception Questionnaire – Revised; identity subscale; EAT= Eating Attitudes Test 26;
DBS= Decisional Balance Scale. Higher scores on the SATAQ, ATSPPHS, CORE, IPQR, EAT, DBS-Benefits, DBS-Burdens, and DBS-Functional avoidance indicate more
internalization of sociocultural beauty norms, more stigma towards counseling, more
general psychological distress, more insight that individual symptoms are related to an
underlying disorder, more disordered eating behavior, more perceived benefits to the
disorder, more perceived drawbacks to the disorder, and a greater awareness of the
function the disorder serves to avoid uncomfortable emotion, respectively.
129
Table 5
Correlations Among Demographics and Constructs
Variable
Age
SATAQ
ATSPPHS
CORE
IPQ-R
Identity
(Insight)
Eating
Attitudes
Test
DBS
DBS
Benefits
Burdens
Age
SATAQ
-.15*
ATSPPHS
-.17*
-.04
CORE
-.07
.18*
-.02
IPQ-R
-.02
.43**
-.07
.23**
EAT
-.03
.17*
-.01
.26**
.35**
Benefits
-.14*
.41**
.07
.32**
.61**
.33**
Burdens
-.01
.29**
-.16*
.47**
.57**
.35**
.58**
Functional
Avoidance
.01
.09
-.05
.42**
.39**
.33**
.44**
DBS
.70**
Note: An * indicates significance at p < .05; An ** indicates significance at p < .001;
SATAQ = Sociocultural Attitudes Towards Appearance Questionnaire; ATSPPHS=
Attitudes Towards Seeking Professional Psychological Help Scale; CORE= Clinical
Outcomes in Routine Evaluation – Short Form; IPQ-R= Illness Perception Questionnaire
– Revised; identity subscale; EAT= Eating Attitudes Test 26; DBS= Decisional Balance
Scale. Higher scores on the SATAQ, ATSPPHS, CORE, IPQ-R, EAT, DBS-Benefits,
DBS-Burdens, and DBS-Functional avoidance indicate more internalization of
sociocultural beauty norms, more stigma towards counseling, more general psychological
distress, more insight that individual symptoms are related to an underlying disorder,
more disordered eating behavior, more perceived benefits to the disorder, more perceived
drawbacks to the disorder, and a greater awareness of the function the disorder serves to
avoid uncomfortable emotion, respectively.
130
Table 6
Analyses with Constructs
Control
variable
Predictor
variable
Moderator
Criterion
variable
Binary
Logistic
Regression
Hierarchical
Regression
Insight x EAT-26
Internalization x EAT-26
CORE-SF EAT-26 Benefits
Benefits x EAT-26
ATSPPHS
Burdens
Burdens x EAT-26
Func. Avoid.
Func. Avoid. x EAT-26
Insight
Insight x EAT-26
Yes/No
Internalization
Internalization x EAT-26 information
CORE-SF EAT-26 Benefits
Benefits x EAT-26
about ED
Burdens
Burdens x EAT-26
and
Func. Avoid.
Func. Avoid. x EAT-26
treatment
Insight
Insight x EAT-26
Internalization
Internalization x EAT-26 Yes/No
CORE-SF EAT-26 Benefits
Benefits x EAT-26
make appt
Burdens
Burdens x EAT-26
for
Func. Avoid.
Func. Avoid. x EAT-26
counseling
Note: CORE-SF = Clinical Outcomes in Routine Evaluation – Short Form; EAT-26 =
Eating Attitudes Test – 26; Func. Avoid. = DBS-Functional Avoidance; ATSPPHS =
Attitudes Towards Seeking Professional Psychological Help Scale. All variables were
standardized.
Binary
Logistic
Regressio
Insight
Interaction term
Internalization
131
Table 7
Hierarchical Multiple Regression Results with IPQ-R (insight) as a Moderator
Step 1
Step 2
Step 3
R
.017
.056
.084
R2
.000
.003
.007
F
.071
.259
.430
p
.790
.855
.787
ΔR2
F
p
.003
.004
.35
.95
.703
.332
Note: N = 249; IPQ-R = Illness Perception Questionnaire – Revised; Step 1 = CORE;
Step 2 = Step1 + Insight + EAT; Step 3 = Step 2 + (EAT x Insight)
Table 8
Binary Logistic Regression Model Results with IPQ-R (insight) as a moderator
Predicted
Variable
Online
ED info
Model
1
2
3
R2
.031
.037
.037
χ2
df
p
7.96*
9.36*
9.51*
1
3
4
.005
.025
.050
Counseling
Cent r
Appointment
1
.011
2.72
1
.099
2
.015
3.83
3
.280
3
.016
3.94
4
.414
Note: N = 249; IPQ-R = Illness Perception Questionnaire – Revised;
An * indicates significance at p < .05; Model 1 = CORE; Model 2 =
CORE+ EAT + Insight; Model 3 = CORE + EAT + Insight +
(EAT x Insight).
Table 9
Binary Logistic Regression Variable Results predicting desire to view online
information about disordered eating with IPQ-R (insight) as a Moderator
Variable
CORE
EAT
IPQ-R
Interaction
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
-.50*
6.18
.013
.61
.41
.90
.09
.14
.709
1.09
.69
1.71
-.26
.23
.258
.77
.49
1.21
-.07
.14
.706
.93
.64
1.36
Note: N = 249; IPQ-R = Illness Perception Questionnaire – Revised; An * indicates
significance at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT +
Insight; Model 3 = CORE + EAT + Insight + (EAT x Insight).
132
Table 10
Binary Logistic Regression Variable Results predicting desire to schedule a
counseling appointment with IPQ-R (insight) as a Moderator
Variable
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
CORE
EAT
IPQ-R
Interaction
-.48
1.52
.218
.61
.29
1.33
-.18
.30
.584
.84
.45
1.58
-.45
.76
.384
.64
.23
1.76
.08
.12
.734
1.09
.67
1.76
Note: N = 249; IPQ-R = Illness Perception Questionnaire – Revised; An * indicates
significance at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT +
Insight; Model 3 = CORE + EAT + Insight + (EAT x Insight).
Table 11
Hierarchical Multiple Regression Results with SATAQ as a Moderator
R
R2
F
p
ΔR2
F
p
Step 1
.017
.000
.071
.790
Step 2
.037
.001
.115
.952
.001
.14
.873
Step 3
.045
.002
.125
.973
.001
.16
.690
Note: N = 249; SATAQ = Sociocultural Attitudes Towards Appearance Questionnaire;
Step 1 = CORE; Step 2 = Step 1 + EAT + SATAQ ; Step 3 = Step 2 +
(EAT x SATAQ)
Table 12
Binary Logistic Regression Model Results with SATAQ as a Moderator
Predicted
Variable
Online
ED info
Counseling
Center
Appointment
Model
1
2
3
R2
χ2
df
p
.031
.033
.033
7.96*
8.32*
8.37
1
3
4
.005
.040
.079
1
.011
2.72
1
.099
2
.013
3.21
3
.361
3
.019
4.90
4
.298
Note: N = 249; SATAQ = Sociocultural Attitudes Towards Appearance
Questionnaire; An * indicates significance at p < .05; Model 1 = CORE;
Model 2 = CORE + EAT + SATAQ; Model 3 = CORE + EAT +SATAQ
+ (EAT x SATAQ)
133
Table 13
Binary Logistic Regression Variable Results predicting desire to view online
information about disordered eating with SATAQ as a Moderator
Variable
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
CORE
EAT
SATAQ
Interaction
-.52*
6.93
.008
.59
.40
.88
.03
.02
.790
1.03
.70
1.51
-.11
.30
.583
.89
.59
1.34
.05
.05
.818
1.05
.69
1.59
Note: N = 249; SATAQ = Sociocultural Attitudes Toward Appearance
Questionnaire; An * indicates significance at p < .05; Model 1 = CORE; Model
2 = CORE+ EAT + SATAQ; Model 3 = CORE + EAT + SATAQ + (EAT x SATAQ).
Table 14
Binary Logistic Regression Variable Results predicting desire to schedule a
counseling appointment with SATAQ as a Moderator
Variable
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
CORE
EAT
SATAQ
Interaction
-.63
2.53
.112
.54
.25
1.16
-.04
.01
.906
.96
.49
1.90
.14
.12
.730
1.15
.53
2.51
.45
1.50
.220
1.57
.76
3.25
Note: N = 249; SATAQ = Sociocultural Attitudes Toward Appearance
Questionnaire; An * indicates significance at p < .05; Model 1 = CORE; Model
2 = CORE+ EAT + SATAQ; Model 3 = CORE + EAT + SATAQ + (EAT x SATAQ).
Table 15
Hierarchical Multiple Regression Results with DBS-Benefits as a Moderator
R
R2
F
p
ΔR2
F
Step 1
.017
.000
.071
.790
Step 2
.084
.007
.579
.629
.007
.83
Step 3
.094
.009
.539
.707
.002
.42
Note: N = 249; DBS = Decisional Balance Scale; Step 1 = CORE; Step 2 =
Step 1 + Benefits + EAT; Step 3 = Step 2 + (EAT x Benefits)
p
.436
.517
134
Table 16
Binary Logistic Regression Model Results with DBS-Benefits as a Moderator
Predicted
Variable
Online
ED info
Model
1
2
3
R2
χ2
df
p
.031
.043
.054
7.96*
11.02*
13.82*
1
3
4
.005
.012
.008
Counseling
Center
Appointment
1
.011
2.72
1
.099
2
.012
3.07
3
.381
3
.016
4.14
4
.388
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE + EAT + DBS-Benefits; Model 3 =
CORE + EAT +DBS-Benefits + (EAT x DBS-Benefits)
Table 17
Binary Logistic Regression Variable Results predicting desire to view online
information about disordered eating with DBS-Benefits as a Moderator
Variable
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
CORE
EAT
DBS-Benefits
Interaction
-.48*
5.38
.020
.62
.41
.93
.02
.01
.940
1.02
.69
1.50
-.46
3.68
.055
.63
.39
1.01
.05
3.04
.081
1.32
.97
1.82
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT + DBS-Benefits; Model 3
= CORE + EAT + DBS-Benefits + (EAT x DBS-Benefits).
Table 18
Binary Logistic Regression Variable Results predicting desire to schedule a
counseling appointment with DBS-Benefits as a Moderator
Variable
CORE
EAT
DBS-Benefits
Interaction
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
-.62
2.38
.123
.54
.25
1.18
-.21
.46
.498
.81
.45
1.48
-.10
.05
.817
.90
.37
2.18
.25
1.09
.297
1.28
.80
2.05
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT + DBS-Benefits;
Model 3 = CORE + EAT + DBS-Benefits + (EAT x DBS-Benefits).
135
Table 19
Regression Results with DBS-Burdens as a Moderator
R
R2
F
p
ΔR2
F
Step 1
.017
.000
.07
.790
Step 2
.179
.032
2.72*
.045
.032
4.04*
Step 3
.200
.040
2.55*
.040
.008
2.03
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Step 1 = CORE; Step 2 = Step 1 + EAT + Burdens; Step 3 = Step 2
+ (EAT x Burdens)
p
.019
.155
Table 20
Contributing Variables of Significant Regression Predicting ATSPPHS
Unstandardized β
Standardized β
t
p
Variable
CORE
.036
.068
.95
.341
EAT
.034
.037
.94
.351
DBS-Burdens
-.105*
.039
-2.65
.009
Interaction
-.040
.028
-1.43
.155
Note: N = 249; CORE = Clinical Outcomes in Routine Evaluation; EAT = Eating
Attitude Test; DBS = Decisional Balance Scale. An * indicates significance at p < .05.
Table 21
Binary Logistic Regression Model Results with DBS-Burdens as a Moderator
Predicted
Variable
Online
ED info
Counseling
Center
Appointment
Model
1
2
3
R2
χ2
df
p
.031
.108
.109
7.96*
28.54*
28.65*
1
3
4
.005
.000
.000
1
.011
2.72
1
.099
2
.047
12.06*
3
.007
3
.052
13.26*
4
.010
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE + EAT + DBS-Burdens;
Model 3 = CORE + EAT +DBS-Burdens + (EAT x DBS-Burdens)
136
Table 22
Binary Logistic Regression Variable Results predicting desire to view online
information about disordered eating with DBS-Burdens as a Moderator
Confidence Interval
Variable
β
Wald
Statistic
p
Exp(B)
Lower
Upper
CORE
EAT
DBS-Burdens
Interaction
-.17
.53
.467
.84
.53
1.34
.43
1.30
.254
1.53
.74
3.18
-1.07* 17.55
<.001
.34
.21
.57
-.07
.11
.746
.93
.61
1.42
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT + DBS-Burdens; Model 3
= CORE + EAT + DBS-Burdens + (EAT x DBS-Burdens).
Table 23
Binary Logistic Regression Variable Results predicting desire to schedule a
counseling appointment with DBS-Burdens as a Moderator
Variable
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
CORE
EAT
DBS-Burdens
Interaction
.01
.00
.975
1.01
.41
2.51
-.38
.64
.425
.69
.27
1.73
-1.64*
7.23
.007
.20
.06
.64
.37
1.55
.213
1.45
.81
2.59
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE+ EAT + DBS-Burdens;
Model 3 = CORE + EAT + DBS-Burdens + (EAT x DBS-Burdens).
Table 24
Regression Results with DBS- Functional Avoidance as a Moderator
R
R2
F
p
ΔR2
F
Step 1
.017
.000
.071
.790
Step 2
.055
.003
.248
.863
.003
.34
Step 3
.055
.003
.185
.946
.000
.00
Note: N = 249; DBS= Decisional Balance Scale; Step 1 = CORE; Step 2 =
Step 1 + EAT + Functional Avoidance; Step 3 = Step 2 + (EAT x Functional
Avoidance)
p
.715
.985
137
Table 25
Binary Logistic Regression Model Results with DBS-Functional Avoidance
as a Moderator
Predicted
Variable
Model
1
2
3
Online
ED info
R2
χ2
df
p
.031
.065
.072
7.96*
16.87*
18.74*
1
3
4
.005
.001
.001
Counseling
Center
Appointment
1
.011
2.72
1
.099
2
.027
6.86
3
.076
3
.032
8.01
4
.091
Note: N = 249; DBS = Decisional Balance Scale; An * indicates significance
at p < .05; Model 1 = CORE; Model 2 = CORE + EAT + DBS-Functional
Avoidance; Model 3 = CORE + EAT +DBS-Functional Avoidance + (EAT x
DBS-Functional Avoidance)
Table 26
Binary Logistic Regression Variable Results predicting desire to view online
information about disordered eating with DBS-Functional Avoidance as a
Moderator
Confidence Interval
Variable
CORE
EAT
DBS-FA
Interaction
β
Wald
Statistic
p
Exp(B)
Lower
Upper
-.30
1.89
.169
.74
.48
1.14
-.02
.01
.924
.98
.63
1.53
-.62* 10.04
.002
.54
.37
.79
.23
1.89
.169
1.26
.91
1.76
Note: N = 249; DBS-FA = Decisional Balance Scale – Functional Avoidance;
An * indicates significance at p < .05; Model 1 = CORE; Model 2 = CORE+
EAT + DBS-Functional Avoidance; Model 3 = CORE + EAT + DBS-Functional
Avoidance + (EAT x DBS-Functional Avoidance).
138
Table 27
Binary Logistic Regression Variable Results predicting desire to schedule a
counseling appointment with DBS-Functional Avoidance as a Moderator
Variable
CORE
EAT
DBS-FA
Interaction
β
Wald
Statistic
p
Exp(B)
Confidence
Interval
Lower
Upper
-.24
.32
.570
.79
.34
1.80
-.34
.84
.360
.71
.35
1.47
-.80*
4.78
.029
.45
.22
.92
.29
1.25
.264
1.33
.81
2.20
Note: N = 249; DBS-FA = Decisional Balance Scale – Functional Avoidance;
An * indicates significance at p < .05; Model 1 = CORE; Model 2 = CORE+
EAT + DBS-Functional Avoidance; Model 3 = CORE + EAT + DBS-Functional
Avoidance + (EAT x DBS-Functional Avoidance).
139
Appendix A
Eating Attitudes Test – 26
Please read each statement below and indicate how often each item applies to you. In
doing so, use the following options
1=Always 2=Usually 3=Often 4 = Sometimes 5 = Rarely 6 = Never
1. Am terrified about being overweight.
2. Avoid eating when I am hungry.
3. Find myself preoccupied with food.
4. Have gone on eating binges where I feel that I may not be able to stop.
5. Cut my food into small pieces.
6. Aware of the calorie content of the foods that I eat.
7. Particularly avoid food with high carbohydrate content (i.e. bread, rice, potatoes).
8. Feel that others would prefer I ate more.
9. Vomit after I have eaten.
10. Feel extremely guilty after eating.
11. Am preoccupied with a desire to be thinner.
12. Think about burning up calories when I exercise.
13. Other people think that I am too thin.
14. Am preoccupied with the thought of having fat on my body.
15. Take longer than others to eat my meals.
16. Avoid foods with sugar in them.
17. Eat diet foods.
18. Feel that food controls my life.
19. Display self-control around food.
20. Feel that others pressure me to eat.
21. Give too much time and thought to food.
22. Feel uncomfortable after eating sweets.
23. Engage in dieting behavior.
24. Like my stomach to be empty.
25. Have the impulse to vomit after meals.
26. Enjoy trying new rich foods.
1. If you had the resources, would you be interested in seeking help for eating or
body image concerns?
140
Appendix B
EMAIL COMMUNICATION (first email)
Title of Study: Help Seeking Behavior and Health Attitudes
Subject Line: Psychology Study: Earn research credit!
We are seeking psychology undergraduate students to participate in a completely online
study designed to better understand your feelings about therapy. Please note that
participation in any part of the study is entirely voluntary and you may stop at any point.
If you agree to participate in this study, you can earn 2 research credits. This is a
completely online study designed to better understand your perceptions of therapy. You
will be responding to a series of questionnaires including items about your general mental
health, thoughts about therapy, and eating. These measures are expected to take approx
90 minutes.
In addition, records identifying participants will be kept confidential to the extent
permitted by applicable laws and regulations and will not be made publicly available.
You must be 18 years of age to participate in this study.
This is a research study. Please take your time in deciding if you would like to
participate. You are encouraged to ask questions at any time during this study. You may
wish to delay participation until your questions have been answered. For further
information about the study, please contact the Principal Investigator, LeAnn R. Mills,
M.S. at [email protected]. You may also contact the supervising faculty member, Dr.
Lisa Larson, at [email protected] or 294-1487. If you have any questions about the
rights of research subjects or research-related injury, please contact the IRB
Administrator, (515) 294-4566, [email protected], or Director, (515) 294-3115, Office of
Research Assurances, Iowa State University, Ames, Iowa 50011.
If you would like to participate, please go to the SONA website [LINK] and sign in using
the following password: XPL947Q. From the SONA site, you can follow the link to the
online survey.
Thank you in advance for your time.
Sincerely,
LeAnn Mills, M.S.
Doctoral Graduate Student
Department of Psychology
Iowa State University
[email protected]
141
REMINDER EMAIL COMMUNICATION (2nd, 3rd, and 4th email)
Title of Study: Help-Seeking Behavior and Health Attitudes
Subject Line: Reminder - Psychology Study: Earn research credit!
This is a reminder that we are seeking psychology undergraduate students to participate
in a completely online study designed to better understand your feelings about therapy.
Please note that participation in any part of the study is entirely voluntary and you may
stop at any point.
If you agree to participate in this study, you can earn 2 research credits. This is a
completely online study designed to better understand your perceptions of therapy. You
will be responding to a series of questionnaires including items about your general mental
health, thoughts about therapy, and eating. These measures are expected to take approx
90 minutes.
In addition, records identifying participants will be kept confidential to the extent
permitted by applicable laws and regulations and will not be made publicly available.
You must be 18 years of age to participate in this study.
This is a research study. Please take your time in deciding if you would like to
participate. You are encouraged to ask questions at any time during this study. You may
wish to delay participation until your questions have been answered. For further
information about the study, please contact the Principal Investigator, LeAnn R. Mills,
M.S. at [email protected]. You may also contact the supervising faculty member, Dr.
Lisa Larson, at [email protected] or 294-1487. If you have any questions about the
rights of research subjects or research-related injury, please contact the IRB
Administrator, (515) 294-4566, [email protected], or Director, (515) 294-3115, Office of
Research Assurances, Iowa State University, Ames, Iowa 50011.
If you would like to participate, please go to the SONA website [LINK] and sign in using
the following password: XPL947Q. From the SONA site, you can follow the link to the
online survey.
Thank you in advance for your time.
Sincerely,
LeAnn Mills, M.S.
Doctoral Graduate Student
Department of Psychology
Iowa State University
[email protected]
142
Appendix C
STUDY POSTING FORM
Ann Schmidt MUST receive a copy of this form before you send an activation request.
PRINCIPAL INVESTIGATOR (Faculty Supervisor): Lisa Larson
RESEARCHERS: LeAnn Mills & Lisa Larson
STUDY NAME & NUMBER: Help Seeking Behavior and Health Attitudes
BRIEF ABSTRACT: This is a completely online study. Participants will be asked to
complete a series of questionnaires about help-seeking, general mental health, eating, and
sociocultural norms.
STUDY DESCRIPTION (Must be exactly as approved by IRB): This is a completely
online study designed to better understand your feelings about therapy. This online
study consists of questions about your attitudes towards seeking help, sociocultural
norms, eating, and general mental health. The study will take 90 minutes or less to
complete and you will receive 2 credits in your psychology class for your participation.
ELIGIBILITY REQUIREMENTS: To be eligible, participants must have participated in
MASS TESTING and must have received an email invitation from this study’s Principle
Investigator.
DURATION (Minimum 50min.): 60-90 minutes
CREDITS: 2
PREPARATION:
IRB APPROVAL CODE:
IRB APPROVAL EXPIRATION:
IS THIS AN ONLINE STUDY? YES
143
Appendix D
INFORMED CONSENT DOCUMENT
Title of Study:
Help Seeking Behavior and Health Attitudes
Investigators:
LeAnn R. Mills, M.S.
Lisa M. Larson, Ph.D.
This is a research study. Please take your time in deciding if you would like to
participate. You are encouraged to ask questions at any time during this study. You may
wish to delay participation until your questions have been answered. For further
information about the study, please contact the Principal Investigator, LeAnn R. Mills,
M.S. at [email protected]. You may also contact the supervising faculty member, Dr.
Lisa Larson, at [email protected] or 294-1487.
INTRODUCTION
The purpose of this study is to examine your feelings about therapy. You are invited to
participate in this study because you are a student in the Psychology department at Iowa
State University and you participated in the Psychology Department’s Mass Testing in
the Fall of 2010, where you completed a measure indicating you might have negative
thoughts or feelings about food or your body.
DESCRIPTION OF PROCEDURES
This is a completely online study designed to better understand your perceptions of
seeking professional psychological help. You will be responding to a set of survey
measures. These surveys ask questions about your thoughts on seeking counseling, eating
behaviors, and thoughts, feelings, and beliefs about issues with your eating and body
image. This should take about 90 minutes and is worth two research credits. If you
choose to leave the study before reaching the end, credits will be prorated (one credit for
45 minutes). Additionally, if you feel uncomfortable answering any questions, you can
skip these questions and still receive credit.
RISKS
We do not anticipate that these procedures will cause you any harm, but if you experience
discomfort you may talk to the investigators about your concerns. You are free to skip
any question that you do not wish to answer or that make you feel uncomfortable. You
are also free at any time to choose to end your participation. There will be no negative
effects if you choose to skip a question or discontinue your participation in the study. If
you choose to end your participation all data collected will be erased.
144
BENEFITS
If you decide to participate in this study there will be no direct benefit to you other than
learning about psychological research from a participant’s perspective. Your participation
in this project may help the researchers develop a better understanding of factors related
to help seeking for negative body image and food concerns.
COSTS AND COMPENSATION
You will not have any costs from participating in this study. In addition, there is no
monetary compensation for your participation. Rather, you will be compensated by
receiving research credit in your undergraduate psychology course for participating in
this study. If you agree to participate in this study, you will earn 2 research credits for
completing the study; these credits will be prorated (1 credit for 45 minutes) if you
choose to leave the study early. If you choose not to participate, you may contact the
Course Information Office (515-294-8065) for alternative research options in order to
earn research credit for your class.
PARTICIPANT RIGHTS
Your participation in this study is completely voluntary and you may refuse to
participate or leave the study at any time. If you decide to not participate in the
study or leave the study early, it will not result in any penalty or loss of benefits to
which you are otherwise entitled. If you choose to leave the study early, credit
earned will be prorated (one credit for 45 minutes).
CONFIDENTIALITY
Records identifying participants will be kept confidential to the extent permitted by
applicable laws and regulations and will not be made publicly available. Only the
researchers on this study will have access to the data, however, auditing departments of
Iowa State University, and the Institutional Review Board (a committee that reviews and
approves human subject research studies) may inspect and/or copy your records for
quality assurance and data analysis. These records may contain private information.
To ensure confidentiality to the extent permitted by law, the following measures will be
taken to protect your privacy including: (a) assigning you a unique code number that will
be used instead of your name; (b) combining your data with the data collected from other
participants so that no individual information will be identifiable (c) all data will stored in
a locked filing cabinet and/or password protected computer and destroyed 5 years after
results are published (d) the online survey site (SurveyMonkey.com) employs multiple
layers of security to protect all data before it is transmitted to the researcher. If the results
are published, your identity will remain confidential.
145
QUESTIONS OR PROBLEMS
You are encouraged to ask questions at any time during this study. For further
information about the study, you can contact the Principal Investigator, LeAnn R. Mills,
M.S. at [email protected]. You may also contact the supervising faculty member, Dr.
Lisa Larson, at [email protected] or 294-1487.
If you have any questions about the rights of research subjects or research-related injury,
please contact the IRB Administrator, (515) 294-4566, [email protected], or Director,
(515) 294-3115, Office for Responsible Research, Iowa State University, Ames, Iowa
50011.
************************************************************************
******
PARTICIPANT SIGNATURE
By clicking “yes” below, you are indicating that you voluntarily agree to participate in
this study, that the study has been explained to you, that you have been given the time to
read the document and that your questions have been satisfactorily answered. You may
wish to print a copy of this informed consent document for your files since this is an
online study.
Do you agree to participate in this study? If you click “yes”, you will continue to the
survey questions. If you select “No,” you will exit the survey.
a) Yes
b) No
146
Appendix E
Some of the questions in the following study are about your feelings about food, exercise,
and your body. Various questions make an assumption that you have issues with food or
your body image. While this assumption may not be true for everyone, research suggests
that MOST people do have some concerns with negative body image or eating.
Please answer all of these questions as honestly as possible.
Demographic Questions
1. Are you…
a. Male
b. Female
c. Prefer not to answer
2. What is your age?
(open question)
3. What is your student ID number? This will ONLY be used to match this data with
the information you filled out in Mass Testing (open question)
4. What is your major? (open question)
147
Appendix F
CORE-SF
IMPORTANT - PLEASE READ THIS FIRST
This form has 25 statements about how you have been OVER THE LAST WEEK.
Please read each statement and think how often you felt that way last week. Then choose
the number that best represents your response.
1 = Not at all 2 = Only occasionally 3 = Sometimes 4 = Often 5 = Most or all of
the time
Over the last week . . .
1. I have felt terribly alone and isolated
2. I have felt tense, anxious, and nervous
3. I have felt that I have someone to turn to for support when needed
4. I have felt totally lacking in energy and enthusiasm
5. I have felt able to cope when things go wrong
6. I have been troubled by aches, pains, or other physical problems
7. Talking to people has felt too much for me
8. Tension and anxiety have prevented me from doing important things
9. I have been happy with the things I have done
10. I have been disturbed by unwanted thoughts and feelings
11. I have felt panic or terror
12. I have had difficulty getting to sleep or staying asleep
13. I have felt warmth or affection for someone
14. My problems have been impossible to put to one side
15. I have been able to do most of the things that I needed to
16. I have felt despairing or hopeless
17. I have felt criticized by other people
18. I have thought I have no friends
19. I have felt unhappy
20. Unwanted images or memories have been distressing me
21. I have been irritable when with other people
22. I have thought I am to blame for my problems and difficulties
23. I have achieved the things I wanted to
24. I have felt humiliated or shamed by other people
148
Appendix G
IPQ-R
YOUR VIEWS ABOUT FOOD, EXERCISE AND YOUR BODY IMAGE
Symptom/Characteristic
Weighing less than is normal for my
age and height
I have experienced this
since I started having
issues with food and
body image
Yes
No
Degree to which I feel
this is related to my food
and body image issues
(circle one number on
scale of 0-4)
Not at all
Related
Very
Related
0____1____2____3____4
Absent/Irregular Menstrual Periods
Yes
No
Restricting food intake to influence
weight/shape
Yes
No
Having eating binges during which I
lose control of my eating
Yes
No
Using laxatives to control my weight
or shape
Yes
Using diuretics to control my weight
or shape
Yes
No
Making myself sick in order to
control/prevent weight gain
Yes
No
Exercising excessively to control my
weight
Yes
No
Avoiding weighing myself
Yes
No
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
No
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
Weighing myself frequently/ several
times a day
Yes
No
Frequently checking my body (e.g.
looking in mirror)
Yes
No
Avoiding looking at my body (e.g.
in a shop window)
Yes
No
Avoiding eating in public
Yes
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
No
0____1____2____3____4
149
Eating in secret
Yes
No
Making strict rules about what I
should/shouldn’t eat
Yes
No
Feeling guilty after eating
Yes
0____1____2____3____4
0____1____2____3____4
No
0____1____2____3____4
Preoccupation with thoughts about
food, eating, or calories
Yes
No
Preoccupation with thoughts about
body shape or weight
Yes
No
A strong desire to loose weight
Yes
No
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
Being afraid of gaining weight or
becoming fat
Yes
No
Feelings of fatness
Yes
No
Feeling dissatisfied with my weight
or shape because I believe I am too
heavy or large
Shape influencing how I feel about
myself as a person
Yes
No
The belief that my eating/exercise
behaviors give me a sense of value
and self-worth
The belief that without my
eating/exercise issues I have no selfidentity
The belief that my eating/exercise
behavior is a means of
communicating how I feel to others
The belief that without my
eating/exercise choices I am all
alone in the world
The belief that my eating/exercise
choices are the only thing I can
depend on/trust
The belief that my eating/exercise
behaviors keep me safe/protects me
Yes
The belief that my eating/exercise
behaviors help me deal with painful
emotions
Yes
0____1____2____3____4
0____1____2____3____4
0____1____2____3____4
Yes
No
0____1____2____3____4
No
0____1____2____3____4
Yes
No
0____1____2____3____4
Yes
No
0____1____2____3____4
Yes
No
0____1____2____3____4
Yes
No
0____1____2____3____4
Yes
No
0____1____2____3____4
No
0____1____2____3____4
150
The belief that my eating/exercise
behaviors give me a sense of control
Yes
No
The belief that my eating/exercise
behavior gives me a sense of
achievement
Yes
No
0____1____2____3____4
0____1____2____3____4
IPQ-R: VIEWS ABOUT YOUR EATING CONCERNS
We are interested in your own personal views of how you see your current concerns
about eating and your body. Please indicate how much you agree or disagree with the
following statements about your food/body image issues by checking the appropriate box.
Strongly
Disagree
1*
2
3
4*
5
6
7
8*
9
10
11
12
13
14
My issues with food and body image will last
a long time
My issues with food and body image are
likely to be permanent rather than temporary
My issues with food and body image will last
for a long time
This issues with food and body image will
pass quickly
I expect to have these issues with food and
body image for the rest of my life
My issues with food and body image are a
serious condition
My issues with food and body image have
major consequences on my life
My issues with food and body image do not
have much effect on my life
My issues with food and body image strongly
affect the way others see me
My issues with food and body image have
serious financial consequences
My issues with food and body image cause
difficulties for those who are close to me
There is a lot I can do to control my issues
with food and body image
What I do can determine whether my issues
with food and body image get better or worse
The course of my food and body image issues
Disagree
Neither
agree
nor
disagree
Agree
Strongly
Agree
151
15*
16
17*
18*
19*
20
21
22
23*
24*
25*
26*
27*
28
29
30
31
32
33
34
35
36*
37
38
depends on me
Nothing I do will affect my issues with food
and body image
I have the power to influence my issues with
food and body image
My actions will have no affect on the
outcome of my issues with food and body
image
My issues with food and body image will
improve in time
There is very little that can be done to
improve my issues with food and body image
Treatment will be effective in curing my
issues with food and body image
The negative effects of my issues with food
and body image can be prevented (avoided)
by treatment
Treatment can control my issues with food
and body image
There is nothing which can help my issues
with food and body image
The symptoms related to my issues with food
and body image are puzzling to me
My issues with food and body image are a
mystery to me
I don’t understand my issues with food and
body image
My issues with food and body image don’t
make any sense to me
I have a clear picture/understanding of my
issues with food and body image
The specific issues I have with food and body
image change a great deal from day to day
My symptoms come and go in cycles
My issues with food and body image is very
unpredictable
I go through cycles in which my issues with
food and body image get better and worse
I get depressed when I think about my issues
with food and body image
When I think about my issues with food and
body image I get upset
My issues with food and body image make
me angry
My issues with food and body image do not
worry me
Having these issues with food and body
image make me feel anxious
My issues with food and body image make
me afraid
152
IPQ-R: CAUSES
We are interested in what you consider may have been the cause of your issues with food
and body image. As people are very different, there is no correct answer for this
question. We are most interested in your views about the causal factors rather than what
others (including doctors or family) may have suggested to you. Below is a list of
possible causes for your issues with food and body image. Please include how much you
agree or disagree that they were causes for you by checking the appropriate box.
Strongly
Disagree
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Disagree
Neither
agree
nor
disagree
Agree
Strongly
Agree
Stress or worry
Heredity – it runs in my family
A germ of virus
Diet or eating habits
Chance or bad luck
Poor medical care in my past
Pollution in the environment
My own behavior
My mental attitude, e.g. thinking about my
life negatively
Family problems or worry caused my eating
disorder
Overwork
My emotional state, e.g. feeling down, lonely,
anxious, empty
Aging
Alcohol
Smoking
Accident or injury
My personality
Altered immunity
In the table below, please list in rank-order the three most important factors that you now
believe caused YOUR issues with food and body image. You may use any of the items
from the box above, or you may have additional ideas of your own.
The most important causes for me :
______________________
______________________
______________________
153
Appendix H
SATAQ-3
Please read each of the following items carefully and indicate the number that best
reflects your agreement with the statement.
Definitely Disagree = 1
Mostly Disagree = 2
Neither Agree Nor Disagree = 3
Mostly Agree = 4
Definitely Agree = 5
1. TV programs are an important source of information about fashion and "being
attractive."
2. I've felt pressure from TV or magazines to lose weight.
3. I do not care if my body looks like the body of people who are on TV.
4. I compare my body to the bodies of people who are on TV.
5. TV commercials are an important source of information about fashion and "being
attractive."
6. I do not feel pressure from TV or magazines to look pretty.
7. I would like my body to look like the models who appear in magazines.
8. I compare my appearance to the appearance of TV and movie stars.
9. Music videos on TV are not an important source of information about fashion and
"being attractive."
10. I've felt pressure from TV and magazines to be thin.
11. I would like my body to look like the people who are in movies.
12. I do not compare my body to the bodies of people who appear in magazines.
13. Magazine articles are not an important source of information about fashion and
"being attractive."
14. I've felt pressure from TV or magazines to have a perfect body.
15. I wish I looked like the models in music videos.
16. I compare my appearance to the appearance of people in magazines.
17. Magazine advertisements are an important source of information about fashion
and "being attractive."
18. I've felt pressure from TV or magazines to diet.
19. I do not wish to look as athletic as the people in magazines.
20. I compare my body to that of people in "good shape."
21. Pictures in magazines are an important source of information about fashion and
"being attractive."
22. I've felt pressure from TV or magazines to exercise.
23. I wish I looked as athletic as sports stars.
24. I compare my body to that of people who are athletic.
25. Movies are an important source of information about fashion and "being
attractive."
26. I've felt pressure from TV or magazines to change my appearance.
154
27. I do not try to look like the people on TV.
28. Movie starts are not an important source of information about fashion and "being
attractive."
29. Famous people are an important source of information about fashion and "being
attractive."
30. 30. I try to look like sports athletes.
155
Appendix I
DBS
Many people say that there are good things and bad things about having issues with
food and body image. Some of the good and bad things that people have talked about are
listed below. Please rate how much each of the following statements apply to you, from
not at all true, to completely true.
Not at
all
true
1. I hate the fact that my issues with food and
body image control my life
2. My issues with food and body image give me
self-control.
3. Because of my issues with food and body
image, I don’t have to deal with intimate adult
1
relationships
4. Because of my issues with food and body
image, I feel guilty a lot of the time
5. Being a low weight makes me feel confident
6. It bothers me that my weight controls my mood
7. My issues with food and body image help me
1
obtain an immediate goal
8. As long as I have my issues with food and body
image, I don’t have to make definite plans for the
future1
9. I don’t it like that my issues with food and
body image keep me from eating out with others
10. Being a low weight makes me feel good about
myself
11. My issues with food and body image are my
way of avoiding deeper, more serious problems1
12. My issues with food and body image make
me moody
13. Being thinner than others makes me feel good
about myself
14. I spend too much time thinking about food,
eating and calories
15. I am fed up with thinking about my weight
and/or shape
16. My issues with food and body image protect
me from the difficulties of adult life1
17. I worry about the effect my issues with food
and body image are having on my health
A
little
true
Moderately
true
Very
true
Completely
true
156
18. My issues with food and body image are my
way of being perfect
19. It bothers me that because of my issues with
food and body image I can’t prepare a meal for
myself
Not at
all
true
20. I am tired of being sick with food and body
image issues
21. My food and body image issues make me feel
accomplished
22. It bothers me that my issues with food and
body image prevent me from sharing my feelings
with others
23. When I focus on eating, shape and weight, I do
not have to deal with painful emotions
24. It bothers me that my issues with food and
body image leaves me with no energy
25. My issues with food and body image allows
me to avoid making decisions
26. It bothers me that my issues with food and
body image keep me from socializing
27. Fitting into small sized clothes makes me feel
good about myself
28. I worry that because of my issues with food
and body image I will not be able to have children
29. Because of my issues with food and body
image, I can avoid my fears about sex and/or my
sexuality
30. I have lost my freedom to my issues with food
and body image
A
little
true
Moderately
true
Very
true
Completely
true
157
Appendix J
158
Appendix K
Thanks for completing these questionnaires. You can expect to receive your 2 research
credits through SONA within the next 1-2 weeks. Please feel free to contact the
researcher at [email protected] with any concerns about credit.
Would you like to be directed to online information to learn more about eating issues?
a. Yes
b. No thanks
c. Prefer not to answer
Would you like to be directed to the University Student Counseling Service website to
sign up for an appointment to talk about your body image, eating, or exercise?
Counseling there is free.
a. Yes
b. No thanks
c. Prefer not to answer
159
Appendix L
www.something-fishy.org
160
Appendix M
SCS Homepage
161
Appendix N
Debriefing Statement
Thank you for your participation. The study you just participated in was designed
to better understand the effects of eating attitudes on help-seeking behavior. Increased
understanding of help-seeking behavior may lead to interventions to help raise awareness
of and reduce negative perceptions of counseling. As mentioned before, all responses
will be kept confidential and identifying information (i.e., names) will be removed at the
end of you participation today. Your data will also be combined with the data of other
participants to further ensure anonymity. These data will be kept in a locked cabinet, in a
locked office.
If you have any concerns about the study you just participated in, please talk to
one of the experimenters. If participation in this study raised concerns about your body
image, thoughts about food, or eating behaviors that you would like to discuss with a
counselor, please visit one of the community resources listed below.
If you have any additional questions about this investigation you may contact the
Principal Investigator, LeAnn R. Mills, M.S. at [email protected]. The Major
Professor, Dr. Lisa Larson, can be contacted at [email protected] or 294-1487.
If you have any questions about the rights of research subjects or research-related
injury, please contact the IRB Administrator, (515) 294-4566, [email protected], or
Director, Office for Responsible Research, (515) 294-3115, 1138 Pearson Hall, Ames, IA
50011.
Community Referrals
Student Counseling Service. 3rd Floor Student Services Bldg. Ames, IA 294-5056
Thielen Student Health Center. Sheldon Ave. Ames, IA, 294-5801.
Couples and Family Therapy Clinic. 4380 Palmer HDFS Bldg, Ames, IA 294-0534
The Richmond Center. 1619 South High Ave, Ames, IA (515) 232-5811
162
Appendix O
IRB Approval of Study